1Laboratory of Watershed Hydrology and Ecology Cold and Arid Regions Environmental and Engineering Research InstituteChinese Academy of Sciences Lanzhou 730000 China2State Key Laboratory of Cryospheric Science Cold and Arid Regions Environmental and Engineering Research InstituteChinese Academy of Sciences Lanzhou 730000 China3Shule River Basin Water Resources Administration of Gansu Province Yumen 735200 China
Correspondence should be addressed to Jinkui Wu jkwulzbaccn
Received 25 December 2014 Revised 20 April 2015 Accepted 28 April 2015
Academic Editor Fengjing Liu
Copyright copy 2015 Jiaxin Zhou et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
The runoff components were identified in the headwater area of Shule River Basin using isotopic and chemical tracing withparticular focus on the temporal variations of catchment sources A total of 95 samples including precipitation groundwater andglacialmeltwater were collected and analyzed for stablewater isotopes (18Oand 2H) andmajor chemical ion parameters (potassiumsodium calcium magnesium sulfate chloride and bicarbonate) Based on the isotope and water chemistry data we appliedend member mixing analysis (EMMA) to identify and quantify the major runoff generating sources and their contributions Thecontributions of groundwater precipitation and glacial meltwater were 667 199 and 134 respectively The study indicatedthat groundwater dominated runoff in the headwater area of Shule River BasinThe roles of glacier meltwater should be remarkablein water resourcemanagement in this basinThe uncertainties of the EMMAmethodwere summarized and estimated via a classicalGaussian error propagation technique Analyses suggested that the uncertainty in the measurement method was less importantthan that in the temporal and spatial variations of tracer concentrationsThe uncertainty was sensitive when the difference betweenmixing components was small Therefore the variation of tracers and the difference of mixing components should be consideredwhen hydrograph separation was applied in the basin
1 Introduction
The quantification of catchment response to rainfall orsnowmelt events in terms of water fluxes and chemicalcomposition is an important issue in catchment hydrologyIn particular during flooding periods different interactingprocesses occur that are spatially distributedwithin the catch-ment [1 2] These processes are defined by physiographiccharacteristics In addition runoff generation depends onthe initial state of the various hydrological reservoirs and onthe characteristics of the hydrological input (precipitation orsnowmelt) Due to these factors it is difficult to identify thedominant runoff generation processes [3]
The water shortage and low use efficiency make chinathirsty and the loss of glacier and wetland in the westernplateau will exaggerate this thirst in the future [4] whilethe same situation happens in other places of the world [5]Therefore it is important to understand the runoff generationmechanism In arid or semiarid regions water is a keyfactor affecting the biomass production A better under-standing of runoff generation processes as well as catch-ment function is important for improved water resourcesmanagement [6]The hydrograph separation technique usingnatural tracers in which different runoff components arequantified according to their chemical signature is a widelyused method for investigating runoff generation processes at
Hindawi Publishing CorporationAdvances in MeteorologyVolume 2015 Article ID 830306 10 pageshttpdxdoiorg1011552015830306
2 Advances in Meteorology
the catchment scale [7] Isotopes were used to quantify theinteraction of different end member in glaciated catchmentsfor longer time periods (monthly) andor larger catchmentareas (gt1000 km2) [8ndash10] Isotope techniques can be easilyand successfully used to study the origin and dynamics ofsurface water and groundwater evaporation of water bodiesand mixing processes between various water sources [11ndash13]To obtain both temporal and spatial origins some investiga-tions using stable isotopes associated with chemical tracershave been undertaken in several different basins [14 15]Hydrochemical tracers such as PH electrical conductivity orthe concentration of different anions and cations [16] havebeen used to determine the origin of runoff componentsIn recent years geochemical methods and environmentalisotope techniques have been used increasingly to determinerunoff components in various catchments under differentenvironmental conditions [17 18] In the arid and semiaridareas a combination of hydrologic and environmental iso-tope methods (18O D) has been proved to be a valuable toolfor studying processes within the water cycle and in isotopehydrology [19ndash21]
One common tool to identify runoff sources and fluxcomponents and calculate their contributions to the streamdischarge is end member mixing analysis (EMMA) [22]EMMA techniques have been applied in varieties of studiesat both the plot and the catchment scales Only few studieshave applied EMMA in the arid or semiarid regions [23] Ithas been applied in many studies to identify end membersat small catchment scales that describe the vertical sequenceof water storages to flow contribution These vertical endmembers are for example rain soil water and groundwater[24] or overland flow soil water and hillslope water Fewerstudies have applied EMMA at larger catchment scales ofhundreds or even thousands of square kilometers [25] Thisapproach is based on threemass conservations one for waterone for isotopic tracer and one for geochemical tracer Itallows separating the relative contribution of the differentcomponents which correspond to different reservoirs ordifferent contributive areasThe use of isotopic tracers allowsseparating the runoff hydrograph into preevent water andevent water while the use of geochemical tracers allowsidentifying the three origins of the runoff components [2627]
There is a clear need to develop predictive capabilitiesrelated to the identification of runoff generating sources inlarge ungauged basins particularly in emergent countriessuch as China [28] Runoff generation and dynamics is animportant issue in watershed and water resource manage-ment On the one hand knowledge about runoff genera-tion processes and flow pathways is crucial for evaluatingthe vulnerability of surface and groundwater system [29]On the other hand such knowledge helps to develop andvalidate hydrological models Since arid and semiarid basinsusually have more severe natural conditions and scarcitiesof observation data the application of isotope techniques incatchment hydrology study seems to be a more economicand helpful tool [30] Understanding hydrological processeswill significantly add our ability to evaluate potential tradeoffs
between social development andwater availabilityWe expectthat scientific results will provide an insight forwater resourceand watershed management in a large-area
In this study we applied the EMMA method to identifyand quantify the major runoff generating sources in a threeend membersrsquo system The objectives of this study are (1) toidentify runoff producing sources using 18O and chloride ionas tracers in the headwater area of the Shule River Basin(2) to investigate the applicability of the EMMA method insemiarid catchments (3) to calculate the contributions ofthe three components of runoff We developed a conceptualhydrograph separation technique namely three componentsrsquomixing model It is based on the steady-state mass balanceequations of water and tracer fluxes in a catchment Inaddition uncertainties analyses were performed for thehydrograph separation
2 Study Areas
The Shule River the third largest inland rivers basin withthe whole catchment area of approximately 1421 times 104 km2is located in the Western Qilian Mountains The upstreamof the Shule River (Figure 1) at 966∘Esim990∘E and 382∘Nsim
400∘Nwith the area of 114times 104 km2 and themean elevationof 3885m is located in the Tianjun Qinghai ProvinceNortheasternmargin of the Tibetan PlateauThemainstreamannual runoff varying significantly during different years is1083 times 108m3 The 53 of the total annual runoff concen-trated between July and September However the runoff dis-tributed unevenly over time The runoff in spring and winteronly makes up for 85 and 10 respectively Accordingto nearly 40 yearsrsquo hydrological data of Changmabao gaugestation the annual runoff of dry years and wet years is 536 times
108m3 and 1507 times 108m3 respectivelyIn the headwater area of Shule River Basin our study area
Gahe at 9649∘Esim9858∘E and 3802∘Nsim3912∘Nwith the area4 096 km2 there are 347 glaciers and the area of glaciers is2945 km2 which accounts for 072 of the headwater area(Figure 2) Glaciers are mainly distributed above elevation4 500m which is located in Shule Nanshan and TuolaiNanshan [31] Annual sunshine time is 3 033ndash3 246 hoursThe mean annual elevation is 4 000sim4 500m The meanannual air temperature is approximately minus5∘C the annualprecipitation is 100ndash300mm andmainly falling betweenMayand September and annual evaporation is about 1 200mm[32] The temperatures of the hottest month (July) and thecoldest month (January) are 75∘C and minus175∘C respectivelyThe study area belongs to the continental arid desert climateregionwhich is characterized by cold drywinters and relativewarm wet summers [33] In the growing season of Maythrough September the plentiful sunshine and rainfall (80of annual total precipitation) allows plants to grow efficiently
The Quaternary sediments comprising diluvial-alluvialaeolian and lacustrine deposits form the main aquifers inthe basin These sediments are enriched in calcite gypsumand mirabilite in parts of the middle reaches and soilsalinization occurs widely in the middle and lower reachesThe depth to the water table is 5ndash10m There formed
Advances in Meteorology 3
(km)
Gahe
Suli
Laohugou
Changmabao
Weather station
15 30 60 900
Hydrological stationRiverWatershed boundary
99∘09984000998400998400E
99∘09984000998400998400E
98∘09984000998400998400E
98∘09984000998400998400E
97∘09984000998400998400E
97∘09984000998400998400E
W EN
S
40∘ 0
998400 0998400998400
N39∘ 0
998400 0998400998400
N
40∘ 0
998400 0998400998400
N39∘ 0
998400 0998400998400
N Elevation (m)High 5763Low 1855
Glacial meltwaterSample point
PrecipitationRiver waterGroundwater
Figure 1 Site map showing the upstream of Shule River catchmentand sampling sites
N
(km)10 20 400
RiverGlacier
ElevationHigh 5763Low 3501
Figure 2 Glacier distribution of Gahe region
a large area of swamp in the source placesThe aquifer systemincludes a thick unconfined zone consisting of coarse-grainedgravel sand and a confined part consisting of medium tofine and silty sand The landscape is characterized by largemountain ranges with steep valleys and gorges interspersedwith relatively level and wide intermountain grassland basins[34]
3 Material and Methods
31 Field Sampling Intensive synoptic sampling was car-ried out between April and September 2009 in Gahe theheadwater area of the Shule River consisting of 95 samples
Precipitation glacial meltwater groundwater and river waterwere sampled The number of four kinds of samples isprecipitation 15 river water 30 groundwater 31 and glacialmeltwater 19 respectively Precipitation glacial meltwatergroundwater and river water were sampled and analyzedfor stable water isotopes (18O and 2H) major ion chemistryparameters as well Samples were collected in polyethylenebottles and filtered through 045mm Millipore membranefor major element analyses Meteorological parameters andhydrology data were measured continuously by means of anautomatic weather station and gauge station
Precipitation samples were collected immediately aftereach precipitation event in order to minimize the alterationof heavy isotopes by evaporation with plastic basin setsRiver water and groundwater samples were collected once aweek Due to the limitations of some nature conditions wecannot get to glaciers distributed around Gahe Hence wecollected the glacial meltwater samples in Laohugou Glaciernumber 12 Due to the background of the same atmosphericcirculation themoisture of the two sites comes from the samesource We considered the substitute is feasible
32 Laboratory Analyses All samples were kept in near-frozen condition and transported to the State Key Laboratoryof Cryospheric Science Cold andArid Regions Environmen-tal and Engineering Research Institute Chinese Academy ofSciences for test
Concentration of anions (Clminus SO42minus) was analyzed
by Ion Chromatography (IC DX-120 Dionex Germany)while HCO3
minus and CO32minus were analyzed by the titration
method Cations K+ Na+ Ca2+ and Mg2+ were analyzedby using Atomic Absorption Spectroscopy (AAS) methodEvery sample value represents the mean of two consecutivemeasurements Measurement errors were less than 1 Thedetection limits of all ions were lower than 01mgL Chlorideion and 18O were finally selected to assess the differentcontributing sources using mass balance equations and endmember mixing diagrams
The 120575D and 12057518O composition of all water samples
were analyzed by Liquid-Water Isotope Analyzer (DLT 100Los Gatos USA) based on off-axis integrated cavity outputspectroscopy (OA-ICOS) Each sample is injected six timesto avoid memory effect between samples The isotopic ratioswere expressed in per mil (permil) units relative to ViennaStandard Mean Ocean Water (V-SMOW)
120575 = (
119877sample
119877SMOWminus 1)times 103 (1)
where 119877 is the ration 18O16O or 2H1H Precision of 120575D and12057518O was plusmn06permil and plusmn02permil respectively
33 Hydrograph Separation Method and Uncertainties Anal-ysis In general hydrograph separations are based on
4 Advances in Meteorology
the steady-state mass balance equations of water and tracerfluxes in a catchment [34] Following are the equations
Then the right side equations are divided by 119876119904
[[[[[[[
[
119876119892
119876119904
119876119901
119876119904
119876119898
119876119904
]]]]]]]
]
=
[[[
[
1 1 1120575119892
120575119901
120575119898
119862119892
119862119901
119862119898
]]]
]
minus1
[[
[
1120575119904
119862119904
]]
]
(4)
where 119876 is the discharge and 119862 and 120575 are the concentrationof tracer chloride ion and 18O respectively Subscripts 119904 119892 119901and 119898 refer to river water groundwater precipitation andglacial meltwater respectively The application of these equa-tions is based on certain assumptions which are discussedfor instance by Hinton et al [35] Buttle [1] or Rodhe [36]
(1) there is a significant difference between the tracerconcentrations of the different components
(2) the tracer concentrations are constant in space andtime or any variations can be accounted for
(3) contributions of an additional component must benegligible or the tracer concentrations must be simi-lar to that of another component
(4) the tracers must mix conservatively
(5) the tracer concentrations of the components are notcollinear
Recent focus of hydrograph separation has been onuncertainty analysis Several approaches are available forcalculating uncertainty Genereux (1998) suggested a generaluncertainty propagation technique using Gaussian errorestimators for two- and three-component separations [37]However an extensive overview of all possible causes ofhydrograph separation uncertainties during different peri-ods of a given event is still lacking A classical Gaussianerror propagation technique was applied to quantify theuncertainty of tracer-based hydrograph separations Thistechnique is generally used in other scientific and engineeringproblems Errors of all separation equation variables areconsidered Assuming that the uncertainty in each variableis independent of the uncertainty in the others the relativeerror 119882
119891of the contribution of a specific runoff component
is related to the uncertainty in each of the variables by thefollowing [37]
119908119910= radic(
120597119910
12059711990911199081199091)
2+ (
120597119910
12059711990921199081199092)
2+ sdot sdot sdot + (
120597119910
120597119909119899
119908119909119899)
2
119908119891119901
= radic[
[
119862119890minus 119862119904
(119862119890minus 119862119901)
2119908119862119901]
]
2
+ [
[
119862119904minus 119862119901
(119862119890minus 119862119901)
2119908119862119890]
]
2
+ [
minus1(119862119890minus 119862119901)
119908119862119904]
2
(5)
where119908 represents the uncertainty in the variable specified inthe subscript 119888 is the concentration of corresponding tracer119890 represents the event water and 119901 represents the preeventwater In the results the relative error is given as percentagevalue
It is demonstrated that large relative uncertainties mustbe considered for the quantification of runoff compo-nents Uncertainties are caused by (1) tracer analysis anddischarge measurement (2) intrastorm variability of 18O(3) elevation effect of 18O and chloride (4) solution ofminerals during runoff formation and (5) general spatialheterogeneity of tracer concentrations The last source oferror was the most significant An investigation on thedominating runoff generation processes in the catchmentbefore a model is set up would reduce such uncertain-ties
4 Results
41 Temporal Variance of Runoff The temporal variance ofrunoff is showed in Figure 3 The runoff showed a significantseasonal variation The runoff varied in the range of 382sim33809m3s with an average of 7157m3s There is a minorpeak in April since the snowmelt peak usually occurs inspring We could see from the figure that most of peakflowswere correspondingwith the big rainfall events betweenJune and September It means the significant increase ofrunoff is the results of precipitation event The averagerunoff is dominated by a snowmelt peak in spring followedby a decline in discharge over the growing season FromJune to September when most rainstorms occur there isa considerable increase in discharge followed by an againdeclining hydrograph until October when the river itselffreezes
Advances in Meteorology 5
0
10
20
30
40
500
50
100
150
200
250
300
350
400
450
50091 10
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
1
Prec
ipita
tion
(mm
)
Julian day
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Runo
ff (m
3s
)
Figure 3 Runoff and rainfall from April to October 2009
42 Isotopic Composition
421 Isotopic Composition of River Water The isotopic com-position of river water during April and September in theGahe station shows a steady variability ranging from minus99 tominus85permil in120575
18Oand fromminus689 tominus582permilin120575D respectivelyThe possibility of differential isotopic evaporation of samplescan be analyzed by comparing the samples to the meteoricwater line which is formed by plotting 120575D against 12057518O [38]The local meteoric water line (LMWL) is commonly used asindicators of water vapor source source of the humidity andkinetic conditions in a number of fields including isotopehydrology [39] The relationship between river water andthe local meteoric water line (LMWL) was displayed bybivariate plot of 120575
18O versus 120575D (Figure 4) According tothe distribution of river water in the space of 120575
18O versus120575D most of the river water sample points were locatedapproaching the local meteoric water line (LMWL) (120575D =81112057518O + 1140 1198772 = 097 119899 = 30) Also the slope of theregression line was fairly close to the multiple-year observedvalues in Northwest China (788) and in Heihe River Basinan inland river basin neighboring the study area (782) [40]
422 Isotopic Signature of Precipitation Groundwater andGlacial Meltwater The isotopic composition of precipitationshows a relative significant variance The values of 120575
18Ofluctuate in minus130simndash83permil and minus951simndash542permil in 120575D Theequation between 120575
18O and 120575D (120575D = 76812057518O + 9291198772
= 097 119899 = 15) (Figure 5(a)) The temporal and spatialvariability of 18O in precipitation are relatively high Thisis caused by fractionation during evapotranspiration andcondensation due to lower saturated vapor pressure of watermolecules containing the heavier 18O isotope than that ofwater molecules containing the lighter 16O isotope As aresult the 120575
18O in precipitation decreases with decreasingair temperature increasing elevation increasing latitudeincreasing distance of vapor transport through the atmo-sphere and increasing precipitation amounts
Figure 4 Stable isotope (120575D and 12057518O) compositions of river water
The isotopic composition of groundwater ranges in minus95simminus67permil (in 120575
18O) and minus682simndash453permil (in 120575D) The equationbetween 120575
18O and 120575D (120575D = 82812057518O + 1074 1198772 = 096119899 = 31) (Figure 5(b)) It is fairly close to that of riverwater The value of 12057518O ranges in minus147simndash123permil and 120575D inminus1050simndash851permil in glacial meltwater The equation between12057518O and 120575D (120575D = 75512057518O + 734 119877
2= 096 119899 =
19) (Figure 5(c)) The stable isotope ratios of hydrogen andoxygen of water samples can provide essential informationabout water dynamics within a given watershed In generalthis is from isotope fractionation by evaporation altitudeeffects and different water sources they received [23] Theslope and the intercept of LMWLwere slightly lower showingdrier and stronger local evaporation conditions Evaporationcaused a differential increase in the 120575D and 120575
18O values ofthe remaining water resulting in a lower slope for the linearrelationship between 120575D and 120575
18O values [41]
43 Temporal Variance of Clminus It can be assumed that mixingprocesses in the catchment determine the isotopic concentra-tion of total runoffHowever the hydrochemical compositionof water is essentially changed as a result of interactions withorganic and inorganicmaterial during its passage through theunsaturated and saturated zones The concentration of Clminusin river water fluctuates in 94sim134mgsdotLminus1 with an averageof 112mgsdotLminus1 The variance of Clminus concentration has muchrelationship with runoff (Figure 6) In spring the springflood caused by snowmelt water makes the soil chemicalions into the river so the concentration of Clminus is relativelyhigh Afterwards with the increase of snowmelt water glaciermeltwater and precipitation the runoff has been showing adifferent amplitudes increaseWith the increase of runoff thedilution effect of ions has also increased so the concentrationof Clminus decreased Although groundwater recharged thedilution effect outweighs the supply effect During Augustand September the runoff has a considerable decrease
Figure 5 Stable isotope (120575D and 12057518O) compositions of precipitation (a) groundwater (b) and glacial meltwater (c)
the concentration of Clminus decreased because of the weakeningof dilution effect
44 Identification of End Members Applying the method ofgeochemical and isotopic tracing in this paper the runoffcharacteristics and hydrological law at the Gahe station areinvestigated at different periods in 2009 We conducted aperformance of principal component analysis (PCA) on theconcentration data of chloride ion The results showed thatthe chemical tracer exhibits conservative behavior whereasisotopes are geographical source tracers and only changecomposition due to slow fractionation processes [42] 18Obelongs to the group of stable environmental isotopes occur-ring naturally in water It has been widely used to separate
storm flow into proportions of event and preevent water [1]As part of the water molecule 18O behaves conservativelythat is the combination of chemical and isotopic tracersallows identifying the origin of water pathways
For the three-component hydrograph separation thechoice of a suitable tracer constellation to explain thechemical changes in discharge during a storm as well asto determine and to identify dominant sources flow pathsand residence times in the catchment becomes increas-ingly important The study suggests that hydrological tracerchlorine concentration and 18O can be used under certainhydrological and lithological conditions A system of alge-braic equations is introduced that enables a three-componenthydrograph separation by using 18O and chlorine These
Advances in Meteorology 7
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
the catchment scale [7] Isotopes were used to quantify theinteraction of different end member in glaciated catchmentsfor longer time periods (monthly) andor larger catchmentareas (gt1000 km2) [8ndash10] Isotope techniques can be easilyand successfully used to study the origin and dynamics ofsurface water and groundwater evaporation of water bodiesand mixing processes between various water sources [11ndash13]To obtain both temporal and spatial origins some investiga-tions using stable isotopes associated with chemical tracershave been undertaken in several different basins [14 15]Hydrochemical tracers such as PH electrical conductivity orthe concentration of different anions and cations [16] havebeen used to determine the origin of runoff componentsIn recent years geochemical methods and environmentalisotope techniques have been used increasingly to determinerunoff components in various catchments under differentenvironmental conditions [17 18] In the arid and semiaridareas a combination of hydrologic and environmental iso-tope methods (18O D) has been proved to be a valuable toolfor studying processes within the water cycle and in isotopehydrology [19ndash21]
One common tool to identify runoff sources and fluxcomponents and calculate their contributions to the streamdischarge is end member mixing analysis (EMMA) [22]EMMA techniques have been applied in varieties of studiesat both the plot and the catchment scales Only few studieshave applied EMMA in the arid or semiarid regions [23] Ithas been applied in many studies to identify end membersat small catchment scales that describe the vertical sequenceof water storages to flow contribution These vertical endmembers are for example rain soil water and groundwater[24] or overland flow soil water and hillslope water Fewerstudies have applied EMMA at larger catchment scales ofhundreds or even thousands of square kilometers [25] Thisapproach is based on threemass conservations one for waterone for isotopic tracer and one for geochemical tracer Itallows separating the relative contribution of the differentcomponents which correspond to different reservoirs ordifferent contributive areasThe use of isotopic tracers allowsseparating the runoff hydrograph into preevent water andevent water while the use of geochemical tracers allowsidentifying the three origins of the runoff components [2627]
There is a clear need to develop predictive capabilitiesrelated to the identification of runoff generating sources inlarge ungauged basins particularly in emergent countriessuch as China [28] Runoff generation and dynamics is animportant issue in watershed and water resource manage-ment On the one hand knowledge about runoff genera-tion processes and flow pathways is crucial for evaluatingthe vulnerability of surface and groundwater system [29]On the other hand such knowledge helps to develop andvalidate hydrological models Since arid and semiarid basinsusually have more severe natural conditions and scarcitiesof observation data the application of isotope techniques incatchment hydrology study seems to be a more economicand helpful tool [30] Understanding hydrological processeswill significantly add our ability to evaluate potential tradeoffs
between social development andwater availabilityWe expectthat scientific results will provide an insight forwater resourceand watershed management in a large-area
In this study we applied the EMMA method to identifyand quantify the major runoff generating sources in a threeend membersrsquo system The objectives of this study are (1) toidentify runoff producing sources using 18O and chloride ionas tracers in the headwater area of the Shule River Basin(2) to investigate the applicability of the EMMA method insemiarid catchments (3) to calculate the contributions ofthe three components of runoff We developed a conceptualhydrograph separation technique namely three componentsrsquomixing model It is based on the steady-state mass balanceequations of water and tracer fluxes in a catchment Inaddition uncertainties analyses were performed for thehydrograph separation
2 Study Areas
The Shule River the third largest inland rivers basin withthe whole catchment area of approximately 1421 times 104 km2is located in the Western Qilian Mountains The upstreamof the Shule River (Figure 1) at 966∘Esim990∘E and 382∘Nsim
400∘Nwith the area of 114times 104 km2 and themean elevationof 3885m is located in the Tianjun Qinghai ProvinceNortheasternmargin of the Tibetan PlateauThemainstreamannual runoff varying significantly during different years is1083 times 108m3 The 53 of the total annual runoff concen-trated between July and September However the runoff dis-tributed unevenly over time The runoff in spring and winteronly makes up for 85 and 10 respectively Accordingto nearly 40 yearsrsquo hydrological data of Changmabao gaugestation the annual runoff of dry years and wet years is 536 times
108m3 and 1507 times 108m3 respectivelyIn the headwater area of Shule River Basin our study area
Gahe at 9649∘Esim9858∘E and 3802∘Nsim3912∘Nwith the area4 096 km2 there are 347 glaciers and the area of glaciers is2945 km2 which accounts for 072 of the headwater area(Figure 2) Glaciers are mainly distributed above elevation4 500m which is located in Shule Nanshan and TuolaiNanshan [31] Annual sunshine time is 3 033ndash3 246 hoursThe mean annual elevation is 4 000sim4 500m The meanannual air temperature is approximately minus5∘C the annualprecipitation is 100ndash300mm andmainly falling betweenMayand September and annual evaporation is about 1 200mm[32] The temperatures of the hottest month (July) and thecoldest month (January) are 75∘C and minus175∘C respectivelyThe study area belongs to the continental arid desert climateregionwhich is characterized by cold drywinters and relativewarm wet summers [33] In the growing season of Maythrough September the plentiful sunshine and rainfall (80of annual total precipitation) allows plants to grow efficiently
The Quaternary sediments comprising diluvial-alluvialaeolian and lacustrine deposits form the main aquifers inthe basin These sediments are enriched in calcite gypsumand mirabilite in parts of the middle reaches and soilsalinization occurs widely in the middle and lower reachesThe depth to the water table is 5ndash10m There formed
Advances in Meteorology 3
(km)
Gahe
Suli
Laohugou
Changmabao
Weather station
15 30 60 900
Hydrological stationRiverWatershed boundary
99∘09984000998400998400E
99∘09984000998400998400E
98∘09984000998400998400E
98∘09984000998400998400E
97∘09984000998400998400E
97∘09984000998400998400E
W EN
S
40∘ 0
998400 0998400998400
N39∘ 0
998400 0998400998400
N
40∘ 0
998400 0998400998400
N39∘ 0
998400 0998400998400
N Elevation (m)High 5763Low 1855
Glacial meltwaterSample point
PrecipitationRiver waterGroundwater
Figure 1 Site map showing the upstream of Shule River catchmentand sampling sites
N
(km)10 20 400
RiverGlacier
ElevationHigh 5763Low 3501
Figure 2 Glacier distribution of Gahe region
a large area of swamp in the source placesThe aquifer systemincludes a thick unconfined zone consisting of coarse-grainedgravel sand and a confined part consisting of medium tofine and silty sand The landscape is characterized by largemountain ranges with steep valleys and gorges interspersedwith relatively level and wide intermountain grassland basins[34]
3 Material and Methods
31 Field Sampling Intensive synoptic sampling was car-ried out between April and September 2009 in Gahe theheadwater area of the Shule River consisting of 95 samples
Precipitation glacial meltwater groundwater and river waterwere sampled The number of four kinds of samples isprecipitation 15 river water 30 groundwater 31 and glacialmeltwater 19 respectively Precipitation glacial meltwatergroundwater and river water were sampled and analyzedfor stable water isotopes (18O and 2H) major ion chemistryparameters as well Samples were collected in polyethylenebottles and filtered through 045mm Millipore membranefor major element analyses Meteorological parameters andhydrology data were measured continuously by means of anautomatic weather station and gauge station
Precipitation samples were collected immediately aftereach precipitation event in order to minimize the alterationof heavy isotopes by evaporation with plastic basin setsRiver water and groundwater samples were collected once aweek Due to the limitations of some nature conditions wecannot get to glaciers distributed around Gahe Hence wecollected the glacial meltwater samples in Laohugou Glaciernumber 12 Due to the background of the same atmosphericcirculation themoisture of the two sites comes from the samesource We considered the substitute is feasible
32 Laboratory Analyses All samples were kept in near-frozen condition and transported to the State Key Laboratoryof Cryospheric Science Cold andArid Regions Environmen-tal and Engineering Research Institute Chinese Academy ofSciences for test
Concentration of anions (Clminus SO42minus) was analyzed
by Ion Chromatography (IC DX-120 Dionex Germany)while HCO3
minus and CO32minus were analyzed by the titration
method Cations K+ Na+ Ca2+ and Mg2+ were analyzedby using Atomic Absorption Spectroscopy (AAS) methodEvery sample value represents the mean of two consecutivemeasurements Measurement errors were less than 1 Thedetection limits of all ions were lower than 01mgL Chlorideion and 18O were finally selected to assess the differentcontributing sources using mass balance equations and endmember mixing diagrams
The 120575D and 12057518O composition of all water samples
were analyzed by Liquid-Water Isotope Analyzer (DLT 100Los Gatos USA) based on off-axis integrated cavity outputspectroscopy (OA-ICOS) Each sample is injected six timesto avoid memory effect between samples The isotopic ratioswere expressed in per mil (permil) units relative to ViennaStandard Mean Ocean Water (V-SMOW)
120575 = (
119877sample
119877SMOWminus 1)times 103 (1)
where 119877 is the ration 18O16O or 2H1H Precision of 120575D and12057518O was plusmn06permil and plusmn02permil respectively
33 Hydrograph Separation Method and Uncertainties Anal-ysis In general hydrograph separations are based on
4 Advances in Meteorology
the steady-state mass balance equations of water and tracerfluxes in a catchment [34] Following are the equations
Then the right side equations are divided by 119876119904
[[[[[[[
[
119876119892
119876119904
119876119901
119876119904
119876119898
119876119904
]]]]]]]
]
=
[[[
[
1 1 1120575119892
120575119901
120575119898
119862119892
119862119901
119862119898
]]]
]
minus1
[[
[
1120575119904
119862119904
]]
]
(4)
where 119876 is the discharge and 119862 and 120575 are the concentrationof tracer chloride ion and 18O respectively Subscripts 119904 119892 119901and 119898 refer to river water groundwater precipitation andglacial meltwater respectively The application of these equa-tions is based on certain assumptions which are discussedfor instance by Hinton et al [35] Buttle [1] or Rodhe [36]
(1) there is a significant difference between the tracerconcentrations of the different components
(2) the tracer concentrations are constant in space andtime or any variations can be accounted for
(3) contributions of an additional component must benegligible or the tracer concentrations must be simi-lar to that of another component
(4) the tracers must mix conservatively
(5) the tracer concentrations of the components are notcollinear
Recent focus of hydrograph separation has been onuncertainty analysis Several approaches are available forcalculating uncertainty Genereux (1998) suggested a generaluncertainty propagation technique using Gaussian errorestimators for two- and three-component separations [37]However an extensive overview of all possible causes ofhydrograph separation uncertainties during different peri-ods of a given event is still lacking A classical Gaussianerror propagation technique was applied to quantify theuncertainty of tracer-based hydrograph separations Thistechnique is generally used in other scientific and engineeringproblems Errors of all separation equation variables areconsidered Assuming that the uncertainty in each variableis independent of the uncertainty in the others the relativeerror 119882
119891of the contribution of a specific runoff component
is related to the uncertainty in each of the variables by thefollowing [37]
119908119910= radic(
120597119910
12059711990911199081199091)
2+ (
120597119910
12059711990921199081199092)
2+ sdot sdot sdot + (
120597119910
120597119909119899
119908119909119899)
2
119908119891119901
= radic[
[
119862119890minus 119862119904
(119862119890minus 119862119901)
2119908119862119901]
]
2
+ [
[
119862119904minus 119862119901
(119862119890minus 119862119901)
2119908119862119890]
]
2
+ [
minus1(119862119890minus 119862119901)
119908119862119904]
2
(5)
where119908 represents the uncertainty in the variable specified inthe subscript 119888 is the concentration of corresponding tracer119890 represents the event water and 119901 represents the preeventwater In the results the relative error is given as percentagevalue
It is demonstrated that large relative uncertainties mustbe considered for the quantification of runoff compo-nents Uncertainties are caused by (1) tracer analysis anddischarge measurement (2) intrastorm variability of 18O(3) elevation effect of 18O and chloride (4) solution ofminerals during runoff formation and (5) general spatialheterogeneity of tracer concentrations The last source oferror was the most significant An investigation on thedominating runoff generation processes in the catchmentbefore a model is set up would reduce such uncertain-ties
4 Results
41 Temporal Variance of Runoff The temporal variance ofrunoff is showed in Figure 3 The runoff showed a significantseasonal variation The runoff varied in the range of 382sim33809m3s with an average of 7157m3s There is a minorpeak in April since the snowmelt peak usually occurs inspring We could see from the figure that most of peakflowswere correspondingwith the big rainfall events betweenJune and September It means the significant increase ofrunoff is the results of precipitation event The averagerunoff is dominated by a snowmelt peak in spring followedby a decline in discharge over the growing season FromJune to September when most rainstorms occur there isa considerable increase in discharge followed by an againdeclining hydrograph until October when the river itselffreezes
Advances in Meteorology 5
0
10
20
30
40
500
50
100
150
200
250
300
350
400
450
50091 10
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
1
Prec
ipita
tion
(mm
)
Julian day
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Runo
ff (m
3s
)
Figure 3 Runoff and rainfall from April to October 2009
42 Isotopic Composition
421 Isotopic Composition of River Water The isotopic com-position of river water during April and September in theGahe station shows a steady variability ranging from minus99 tominus85permil in120575
18Oand fromminus689 tominus582permilin120575D respectivelyThe possibility of differential isotopic evaporation of samplescan be analyzed by comparing the samples to the meteoricwater line which is formed by plotting 120575D against 12057518O [38]The local meteoric water line (LMWL) is commonly used asindicators of water vapor source source of the humidity andkinetic conditions in a number of fields including isotopehydrology [39] The relationship between river water andthe local meteoric water line (LMWL) was displayed bybivariate plot of 120575
18O versus 120575D (Figure 4) According tothe distribution of river water in the space of 120575
18O versus120575D most of the river water sample points were locatedapproaching the local meteoric water line (LMWL) (120575D =81112057518O + 1140 1198772 = 097 119899 = 30) Also the slope of theregression line was fairly close to the multiple-year observedvalues in Northwest China (788) and in Heihe River Basinan inland river basin neighboring the study area (782) [40]
422 Isotopic Signature of Precipitation Groundwater andGlacial Meltwater The isotopic composition of precipitationshows a relative significant variance The values of 120575
18Ofluctuate in minus130simndash83permil and minus951simndash542permil in 120575D Theequation between 120575
18O and 120575D (120575D = 76812057518O + 9291198772
= 097 119899 = 15) (Figure 5(a)) The temporal and spatialvariability of 18O in precipitation are relatively high Thisis caused by fractionation during evapotranspiration andcondensation due to lower saturated vapor pressure of watermolecules containing the heavier 18O isotope than that ofwater molecules containing the lighter 16O isotope As aresult the 120575
18O in precipitation decreases with decreasingair temperature increasing elevation increasing latitudeincreasing distance of vapor transport through the atmo-sphere and increasing precipitation amounts
Figure 4 Stable isotope (120575D and 12057518O) compositions of river water
The isotopic composition of groundwater ranges in minus95simminus67permil (in 120575
18O) and minus682simndash453permil (in 120575D) The equationbetween 120575
18O and 120575D (120575D = 82812057518O + 1074 1198772 = 096119899 = 31) (Figure 5(b)) It is fairly close to that of riverwater The value of 12057518O ranges in minus147simndash123permil and 120575D inminus1050simndash851permil in glacial meltwater The equation between12057518O and 120575D (120575D = 75512057518O + 734 119877
2= 096 119899 =
19) (Figure 5(c)) The stable isotope ratios of hydrogen andoxygen of water samples can provide essential informationabout water dynamics within a given watershed In generalthis is from isotope fractionation by evaporation altitudeeffects and different water sources they received [23] Theslope and the intercept of LMWLwere slightly lower showingdrier and stronger local evaporation conditions Evaporationcaused a differential increase in the 120575D and 120575
18O values ofthe remaining water resulting in a lower slope for the linearrelationship between 120575D and 120575
18O values [41]
43 Temporal Variance of Clminus It can be assumed that mixingprocesses in the catchment determine the isotopic concentra-tion of total runoffHowever the hydrochemical compositionof water is essentially changed as a result of interactions withorganic and inorganicmaterial during its passage through theunsaturated and saturated zones The concentration of Clminusin river water fluctuates in 94sim134mgsdotLminus1 with an averageof 112mgsdotLminus1 The variance of Clminus concentration has muchrelationship with runoff (Figure 6) In spring the springflood caused by snowmelt water makes the soil chemicalions into the river so the concentration of Clminus is relativelyhigh Afterwards with the increase of snowmelt water glaciermeltwater and precipitation the runoff has been showing adifferent amplitudes increaseWith the increase of runoff thedilution effect of ions has also increased so the concentrationof Clminus decreased Although groundwater recharged thedilution effect outweighs the supply effect During Augustand September the runoff has a considerable decrease
Figure 5 Stable isotope (120575D and 12057518O) compositions of precipitation (a) groundwater (b) and glacial meltwater (c)
the concentration of Clminus decreased because of the weakeningof dilution effect
44 Identification of End Members Applying the method ofgeochemical and isotopic tracing in this paper the runoffcharacteristics and hydrological law at the Gahe station areinvestigated at different periods in 2009 We conducted aperformance of principal component analysis (PCA) on theconcentration data of chloride ion The results showed thatthe chemical tracer exhibits conservative behavior whereasisotopes are geographical source tracers and only changecomposition due to slow fractionation processes [42] 18Obelongs to the group of stable environmental isotopes occur-ring naturally in water It has been widely used to separate
storm flow into proportions of event and preevent water [1]As part of the water molecule 18O behaves conservativelythat is the combination of chemical and isotopic tracersallows identifying the origin of water pathways
For the three-component hydrograph separation thechoice of a suitable tracer constellation to explain thechemical changes in discharge during a storm as well asto determine and to identify dominant sources flow pathsand residence times in the catchment becomes increas-ingly important The study suggests that hydrological tracerchlorine concentration and 18O can be used under certainhydrological and lithological conditions A system of alge-braic equations is introduced that enables a three-componenthydrograph separation by using 18O and chlorine These
Advances in Meteorology 7
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
Figure 1 Site map showing the upstream of Shule River catchmentand sampling sites
N
(km)10 20 400
RiverGlacier
ElevationHigh 5763Low 3501
Figure 2 Glacier distribution of Gahe region
a large area of swamp in the source placesThe aquifer systemincludes a thick unconfined zone consisting of coarse-grainedgravel sand and a confined part consisting of medium tofine and silty sand The landscape is characterized by largemountain ranges with steep valleys and gorges interspersedwith relatively level and wide intermountain grassland basins[34]
3 Material and Methods
31 Field Sampling Intensive synoptic sampling was car-ried out between April and September 2009 in Gahe theheadwater area of the Shule River consisting of 95 samples
Precipitation glacial meltwater groundwater and river waterwere sampled The number of four kinds of samples isprecipitation 15 river water 30 groundwater 31 and glacialmeltwater 19 respectively Precipitation glacial meltwatergroundwater and river water were sampled and analyzedfor stable water isotopes (18O and 2H) major ion chemistryparameters as well Samples were collected in polyethylenebottles and filtered through 045mm Millipore membranefor major element analyses Meteorological parameters andhydrology data were measured continuously by means of anautomatic weather station and gauge station
Precipitation samples were collected immediately aftereach precipitation event in order to minimize the alterationof heavy isotopes by evaporation with plastic basin setsRiver water and groundwater samples were collected once aweek Due to the limitations of some nature conditions wecannot get to glaciers distributed around Gahe Hence wecollected the glacial meltwater samples in Laohugou Glaciernumber 12 Due to the background of the same atmosphericcirculation themoisture of the two sites comes from the samesource We considered the substitute is feasible
32 Laboratory Analyses All samples were kept in near-frozen condition and transported to the State Key Laboratoryof Cryospheric Science Cold andArid Regions Environmen-tal and Engineering Research Institute Chinese Academy ofSciences for test
Concentration of anions (Clminus SO42minus) was analyzed
by Ion Chromatography (IC DX-120 Dionex Germany)while HCO3
minus and CO32minus were analyzed by the titration
method Cations K+ Na+ Ca2+ and Mg2+ were analyzedby using Atomic Absorption Spectroscopy (AAS) methodEvery sample value represents the mean of two consecutivemeasurements Measurement errors were less than 1 Thedetection limits of all ions were lower than 01mgL Chlorideion and 18O were finally selected to assess the differentcontributing sources using mass balance equations and endmember mixing diagrams
The 120575D and 12057518O composition of all water samples
were analyzed by Liquid-Water Isotope Analyzer (DLT 100Los Gatos USA) based on off-axis integrated cavity outputspectroscopy (OA-ICOS) Each sample is injected six timesto avoid memory effect between samples The isotopic ratioswere expressed in per mil (permil) units relative to ViennaStandard Mean Ocean Water (V-SMOW)
120575 = (
119877sample
119877SMOWminus 1)times 103 (1)
where 119877 is the ration 18O16O or 2H1H Precision of 120575D and12057518O was plusmn06permil and plusmn02permil respectively
33 Hydrograph Separation Method and Uncertainties Anal-ysis In general hydrograph separations are based on
4 Advances in Meteorology
the steady-state mass balance equations of water and tracerfluxes in a catchment [34] Following are the equations
Then the right side equations are divided by 119876119904
[[[[[[[
[
119876119892
119876119904
119876119901
119876119904
119876119898
119876119904
]]]]]]]
]
=
[[[
[
1 1 1120575119892
120575119901
120575119898
119862119892
119862119901
119862119898
]]]
]
minus1
[[
[
1120575119904
119862119904
]]
]
(4)
where 119876 is the discharge and 119862 and 120575 are the concentrationof tracer chloride ion and 18O respectively Subscripts 119904 119892 119901and 119898 refer to river water groundwater precipitation andglacial meltwater respectively The application of these equa-tions is based on certain assumptions which are discussedfor instance by Hinton et al [35] Buttle [1] or Rodhe [36]
(1) there is a significant difference between the tracerconcentrations of the different components
(2) the tracer concentrations are constant in space andtime or any variations can be accounted for
(3) contributions of an additional component must benegligible or the tracer concentrations must be simi-lar to that of another component
(4) the tracers must mix conservatively
(5) the tracer concentrations of the components are notcollinear
Recent focus of hydrograph separation has been onuncertainty analysis Several approaches are available forcalculating uncertainty Genereux (1998) suggested a generaluncertainty propagation technique using Gaussian errorestimators for two- and three-component separations [37]However an extensive overview of all possible causes ofhydrograph separation uncertainties during different peri-ods of a given event is still lacking A classical Gaussianerror propagation technique was applied to quantify theuncertainty of tracer-based hydrograph separations Thistechnique is generally used in other scientific and engineeringproblems Errors of all separation equation variables areconsidered Assuming that the uncertainty in each variableis independent of the uncertainty in the others the relativeerror 119882
119891of the contribution of a specific runoff component
is related to the uncertainty in each of the variables by thefollowing [37]
119908119910= radic(
120597119910
12059711990911199081199091)
2+ (
120597119910
12059711990921199081199092)
2+ sdot sdot sdot + (
120597119910
120597119909119899
119908119909119899)
2
119908119891119901
= radic[
[
119862119890minus 119862119904
(119862119890minus 119862119901)
2119908119862119901]
]
2
+ [
[
119862119904minus 119862119901
(119862119890minus 119862119901)
2119908119862119890]
]
2
+ [
minus1(119862119890minus 119862119901)
119908119862119904]
2
(5)
where119908 represents the uncertainty in the variable specified inthe subscript 119888 is the concentration of corresponding tracer119890 represents the event water and 119901 represents the preeventwater In the results the relative error is given as percentagevalue
It is demonstrated that large relative uncertainties mustbe considered for the quantification of runoff compo-nents Uncertainties are caused by (1) tracer analysis anddischarge measurement (2) intrastorm variability of 18O(3) elevation effect of 18O and chloride (4) solution ofminerals during runoff formation and (5) general spatialheterogeneity of tracer concentrations The last source oferror was the most significant An investigation on thedominating runoff generation processes in the catchmentbefore a model is set up would reduce such uncertain-ties
4 Results
41 Temporal Variance of Runoff The temporal variance ofrunoff is showed in Figure 3 The runoff showed a significantseasonal variation The runoff varied in the range of 382sim33809m3s with an average of 7157m3s There is a minorpeak in April since the snowmelt peak usually occurs inspring We could see from the figure that most of peakflowswere correspondingwith the big rainfall events betweenJune and September It means the significant increase ofrunoff is the results of precipitation event The averagerunoff is dominated by a snowmelt peak in spring followedby a decline in discharge over the growing season FromJune to September when most rainstorms occur there isa considerable increase in discharge followed by an againdeclining hydrograph until October when the river itselffreezes
Advances in Meteorology 5
0
10
20
30
40
500
50
100
150
200
250
300
350
400
450
50091 10
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
1
Prec
ipita
tion
(mm
)
Julian day
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Runo
ff (m
3s
)
Figure 3 Runoff and rainfall from April to October 2009
42 Isotopic Composition
421 Isotopic Composition of River Water The isotopic com-position of river water during April and September in theGahe station shows a steady variability ranging from minus99 tominus85permil in120575
18Oand fromminus689 tominus582permilin120575D respectivelyThe possibility of differential isotopic evaporation of samplescan be analyzed by comparing the samples to the meteoricwater line which is formed by plotting 120575D against 12057518O [38]The local meteoric water line (LMWL) is commonly used asindicators of water vapor source source of the humidity andkinetic conditions in a number of fields including isotopehydrology [39] The relationship between river water andthe local meteoric water line (LMWL) was displayed bybivariate plot of 120575
18O versus 120575D (Figure 4) According tothe distribution of river water in the space of 120575
18O versus120575D most of the river water sample points were locatedapproaching the local meteoric water line (LMWL) (120575D =81112057518O + 1140 1198772 = 097 119899 = 30) Also the slope of theregression line was fairly close to the multiple-year observedvalues in Northwest China (788) and in Heihe River Basinan inland river basin neighboring the study area (782) [40]
422 Isotopic Signature of Precipitation Groundwater andGlacial Meltwater The isotopic composition of precipitationshows a relative significant variance The values of 120575
18Ofluctuate in minus130simndash83permil and minus951simndash542permil in 120575D Theequation between 120575
18O and 120575D (120575D = 76812057518O + 9291198772
= 097 119899 = 15) (Figure 5(a)) The temporal and spatialvariability of 18O in precipitation are relatively high Thisis caused by fractionation during evapotranspiration andcondensation due to lower saturated vapor pressure of watermolecules containing the heavier 18O isotope than that ofwater molecules containing the lighter 16O isotope As aresult the 120575
18O in precipitation decreases with decreasingair temperature increasing elevation increasing latitudeincreasing distance of vapor transport through the atmo-sphere and increasing precipitation amounts
Figure 4 Stable isotope (120575D and 12057518O) compositions of river water
The isotopic composition of groundwater ranges in minus95simminus67permil (in 120575
18O) and minus682simndash453permil (in 120575D) The equationbetween 120575
18O and 120575D (120575D = 82812057518O + 1074 1198772 = 096119899 = 31) (Figure 5(b)) It is fairly close to that of riverwater The value of 12057518O ranges in minus147simndash123permil and 120575D inminus1050simndash851permil in glacial meltwater The equation between12057518O and 120575D (120575D = 75512057518O + 734 119877
2= 096 119899 =
19) (Figure 5(c)) The stable isotope ratios of hydrogen andoxygen of water samples can provide essential informationabout water dynamics within a given watershed In generalthis is from isotope fractionation by evaporation altitudeeffects and different water sources they received [23] Theslope and the intercept of LMWLwere slightly lower showingdrier and stronger local evaporation conditions Evaporationcaused a differential increase in the 120575D and 120575
18O values ofthe remaining water resulting in a lower slope for the linearrelationship between 120575D and 120575
18O values [41]
43 Temporal Variance of Clminus It can be assumed that mixingprocesses in the catchment determine the isotopic concentra-tion of total runoffHowever the hydrochemical compositionof water is essentially changed as a result of interactions withorganic and inorganicmaterial during its passage through theunsaturated and saturated zones The concentration of Clminusin river water fluctuates in 94sim134mgsdotLminus1 with an averageof 112mgsdotLminus1 The variance of Clminus concentration has muchrelationship with runoff (Figure 6) In spring the springflood caused by snowmelt water makes the soil chemicalions into the river so the concentration of Clminus is relativelyhigh Afterwards with the increase of snowmelt water glaciermeltwater and precipitation the runoff has been showing adifferent amplitudes increaseWith the increase of runoff thedilution effect of ions has also increased so the concentrationof Clminus decreased Although groundwater recharged thedilution effect outweighs the supply effect During Augustand September the runoff has a considerable decrease
Figure 5 Stable isotope (120575D and 12057518O) compositions of precipitation (a) groundwater (b) and glacial meltwater (c)
the concentration of Clminus decreased because of the weakeningof dilution effect
44 Identification of End Members Applying the method ofgeochemical and isotopic tracing in this paper the runoffcharacteristics and hydrological law at the Gahe station areinvestigated at different periods in 2009 We conducted aperformance of principal component analysis (PCA) on theconcentration data of chloride ion The results showed thatthe chemical tracer exhibits conservative behavior whereasisotopes are geographical source tracers and only changecomposition due to slow fractionation processes [42] 18Obelongs to the group of stable environmental isotopes occur-ring naturally in water It has been widely used to separate
storm flow into proportions of event and preevent water [1]As part of the water molecule 18O behaves conservativelythat is the combination of chemical and isotopic tracersallows identifying the origin of water pathways
For the three-component hydrograph separation thechoice of a suitable tracer constellation to explain thechemical changes in discharge during a storm as well asto determine and to identify dominant sources flow pathsand residence times in the catchment becomes increas-ingly important The study suggests that hydrological tracerchlorine concentration and 18O can be used under certainhydrological and lithological conditions A system of alge-braic equations is introduced that enables a three-componenthydrograph separation by using 18O and chlorine These
Advances in Meteorology 7
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
Then the right side equations are divided by 119876119904
[[[[[[[
[
119876119892
119876119904
119876119901
119876119904
119876119898
119876119904
]]]]]]]
]
=
[[[
[
1 1 1120575119892
120575119901
120575119898
119862119892
119862119901
119862119898
]]]
]
minus1
[[
[
1120575119904
119862119904
]]
]
(4)
where 119876 is the discharge and 119862 and 120575 are the concentrationof tracer chloride ion and 18O respectively Subscripts 119904 119892 119901and 119898 refer to river water groundwater precipitation andglacial meltwater respectively The application of these equa-tions is based on certain assumptions which are discussedfor instance by Hinton et al [35] Buttle [1] or Rodhe [36]
(1) there is a significant difference between the tracerconcentrations of the different components
(2) the tracer concentrations are constant in space andtime or any variations can be accounted for
(3) contributions of an additional component must benegligible or the tracer concentrations must be simi-lar to that of another component
(4) the tracers must mix conservatively
(5) the tracer concentrations of the components are notcollinear
Recent focus of hydrograph separation has been onuncertainty analysis Several approaches are available forcalculating uncertainty Genereux (1998) suggested a generaluncertainty propagation technique using Gaussian errorestimators for two- and three-component separations [37]However an extensive overview of all possible causes ofhydrograph separation uncertainties during different peri-ods of a given event is still lacking A classical Gaussianerror propagation technique was applied to quantify theuncertainty of tracer-based hydrograph separations Thistechnique is generally used in other scientific and engineeringproblems Errors of all separation equation variables areconsidered Assuming that the uncertainty in each variableis independent of the uncertainty in the others the relativeerror 119882
119891of the contribution of a specific runoff component
is related to the uncertainty in each of the variables by thefollowing [37]
119908119910= radic(
120597119910
12059711990911199081199091)
2+ (
120597119910
12059711990921199081199092)
2+ sdot sdot sdot + (
120597119910
120597119909119899
119908119909119899)
2
119908119891119901
= radic[
[
119862119890minus 119862119904
(119862119890minus 119862119901)
2119908119862119901]
]
2
+ [
[
119862119904minus 119862119901
(119862119890minus 119862119901)
2119908119862119890]
]
2
+ [
minus1(119862119890minus 119862119901)
119908119862119904]
2
(5)
where119908 represents the uncertainty in the variable specified inthe subscript 119888 is the concentration of corresponding tracer119890 represents the event water and 119901 represents the preeventwater In the results the relative error is given as percentagevalue
It is demonstrated that large relative uncertainties mustbe considered for the quantification of runoff compo-nents Uncertainties are caused by (1) tracer analysis anddischarge measurement (2) intrastorm variability of 18O(3) elevation effect of 18O and chloride (4) solution ofminerals during runoff formation and (5) general spatialheterogeneity of tracer concentrations The last source oferror was the most significant An investigation on thedominating runoff generation processes in the catchmentbefore a model is set up would reduce such uncertain-ties
4 Results
41 Temporal Variance of Runoff The temporal variance ofrunoff is showed in Figure 3 The runoff showed a significantseasonal variation The runoff varied in the range of 382sim33809m3s with an average of 7157m3s There is a minorpeak in April since the snowmelt peak usually occurs inspring We could see from the figure that most of peakflowswere correspondingwith the big rainfall events betweenJune and September It means the significant increase ofrunoff is the results of precipitation event The averagerunoff is dominated by a snowmelt peak in spring followedby a decline in discharge over the growing season FromJune to September when most rainstorms occur there isa considerable increase in discharge followed by an againdeclining hydrograph until October when the river itselffreezes
Advances in Meteorology 5
0
10
20
30
40
500
50
100
150
200
250
300
350
400
450
50091 10
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
1
Prec
ipita
tion
(mm
)
Julian day
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Runo
ff (m
3s
)
Figure 3 Runoff and rainfall from April to October 2009
42 Isotopic Composition
421 Isotopic Composition of River Water The isotopic com-position of river water during April and September in theGahe station shows a steady variability ranging from minus99 tominus85permil in120575
18Oand fromminus689 tominus582permilin120575D respectivelyThe possibility of differential isotopic evaporation of samplescan be analyzed by comparing the samples to the meteoricwater line which is formed by plotting 120575D against 12057518O [38]The local meteoric water line (LMWL) is commonly used asindicators of water vapor source source of the humidity andkinetic conditions in a number of fields including isotopehydrology [39] The relationship between river water andthe local meteoric water line (LMWL) was displayed bybivariate plot of 120575
18O versus 120575D (Figure 4) According tothe distribution of river water in the space of 120575
18O versus120575D most of the river water sample points were locatedapproaching the local meteoric water line (LMWL) (120575D =81112057518O + 1140 1198772 = 097 119899 = 30) Also the slope of theregression line was fairly close to the multiple-year observedvalues in Northwest China (788) and in Heihe River Basinan inland river basin neighboring the study area (782) [40]
422 Isotopic Signature of Precipitation Groundwater andGlacial Meltwater The isotopic composition of precipitationshows a relative significant variance The values of 120575
18Ofluctuate in minus130simndash83permil and minus951simndash542permil in 120575D Theequation between 120575
18O and 120575D (120575D = 76812057518O + 9291198772
= 097 119899 = 15) (Figure 5(a)) The temporal and spatialvariability of 18O in precipitation are relatively high Thisis caused by fractionation during evapotranspiration andcondensation due to lower saturated vapor pressure of watermolecules containing the heavier 18O isotope than that ofwater molecules containing the lighter 16O isotope As aresult the 120575
18O in precipitation decreases with decreasingair temperature increasing elevation increasing latitudeincreasing distance of vapor transport through the atmo-sphere and increasing precipitation amounts
Figure 4 Stable isotope (120575D and 12057518O) compositions of river water
The isotopic composition of groundwater ranges in minus95simminus67permil (in 120575
18O) and minus682simndash453permil (in 120575D) The equationbetween 120575
18O and 120575D (120575D = 82812057518O + 1074 1198772 = 096119899 = 31) (Figure 5(b)) It is fairly close to that of riverwater The value of 12057518O ranges in minus147simndash123permil and 120575D inminus1050simndash851permil in glacial meltwater The equation between12057518O and 120575D (120575D = 75512057518O + 734 119877
2= 096 119899 =
19) (Figure 5(c)) The stable isotope ratios of hydrogen andoxygen of water samples can provide essential informationabout water dynamics within a given watershed In generalthis is from isotope fractionation by evaporation altitudeeffects and different water sources they received [23] Theslope and the intercept of LMWLwere slightly lower showingdrier and stronger local evaporation conditions Evaporationcaused a differential increase in the 120575D and 120575
18O values ofthe remaining water resulting in a lower slope for the linearrelationship between 120575D and 120575
18O values [41]
43 Temporal Variance of Clminus It can be assumed that mixingprocesses in the catchment determine the isotopic concentra-tion of total runoffHowever the hydrochemical compositionof water is essentially changed as a result of interactions withorganic and inorganicmaterial during its passage through theunsaturated and saturated zones The concentration of Clminusin river water fluctuates in 94sim134mgsdotLminus1 with an averageof 112mgsdotLminus1 The variance of Clminus concentration has muchrelationship with runoff (Figure 6) In spring the springflood caused by snowmelt water makes the soil chemicalions into the river so the concentration of Clminus is relativelyhigh Afterwards with the increase of snowmelt water glaciermeltwater and precipitation the runoff has been showing adifferent amplitudes increaseWith the increase of runoff thedilution effect of ions has also increased so the concentrationof Clminus decreased Although groundwater recharged thedilution effect outweighs the supply effect During Augustand September the runoff has a considerable decrease
Figure 5 Stable isotope (120575D and 12057518O) compositions of precipitation (a) groundwater (b) and glacial meltwater (c)
the concentration of Clminus decreased because of the weakeningof dilution effect
44 Identification of End Members Applying the method ofgeochemical and isotopic tracing in this paper the runoffcharacteristics and hydrological law at the Gahe station areinvestigated at different periods in 2009 We conducted aperformance of principal component analysis (PCA) on theconcentration data of chloride ion The results showed thatthe chemical tracer exhibits conservative behavior whereasisotopes are geographical source tracers and only changecomposition due to slow fractionation processes [42] 18Obelongs to the group of stable environmental isotopes occur-ring naturally in water It has been widely used to separate
storm flow into proportions of event and preevent water [1]As part of the water molecule 18O behaves conservativelythat is the combination of chemical and isotopic tracersallows identifying the origin of water pathways
For the three-component hydrograph separation thechoice of a suitable tracer constellation to explain thechemical changes in discharge during a storm as well asto determine and to identify dominant sources flow pathsand residence times in the catchment becomes increas-ingly important The study suggests that hydrological tracerchlorine concentration and 18O can be used under certainhydrological and lithological conditions A system of alge-braic equations is introduced that enables a three-componenthydrograph separation by using 18O and chlorine These
Advances in Meteorology 7
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
Figure 3 Runoff and rainfall from April to October 2009
42 Isotopic Composition
421 Isotopic Composition of River Water The isotopic com-position of river water during April and September in theGahe station shows a steady variability ranging from minus99 tominus85permil in120575
18Oand fromminus689 tominus582permilin120575D respectivelyThe possibility of differential isotopic evaporation of samplescan be analyzed by comparing the samples to the meteoricwater line which is formed by plotting 120575D against 12057518O [38]The local meteoric water line (LMWL) is commonly used asindicators of water vapor source source of the humidity andkinetic conditions in a number of fields including isotopehydrology [39] The relationship between river water andthe local meteoric water line (LMWL) was displayed bybivariate plot of 120575
18O versus 120575D (Figure 4) According tothe distribution of river water in the space of 120575
18O versus120575D most of the river water sample points were locatedapproaching the local meteoric water line (LMWL) (120575D =81112057518O + 1140 1198772 = 097 119899 = 30) Also the slope of theregression line was fairly close to the multiple-year observedvalues in Northwest China (788) and in Heihe River Basinan inland river basin neighboring the study area (782) [40]
422 Isotopic Signature of Precipitation Groundwater andGlacial Meltwater The isotopic composition of precipitationshows a relative significant variance The values of 120575
18Ofluctuate in minus130simndash83permil and minus951simndash542permil in 120575D Theequation between 120575
18O and 120575D (120575D = 76812057518O + 9291198772
= 097 119899 = 15) (Figure 5(a)) The temporal and spatialvariability of 18O in precipitation are relatively high Thisis caused by fractionation during evapotranspiration andcondensation due to lower saturated vapor pressure of watermolecules containing the heavier 18O isotope than that ofwater molecules containing the lighter 16O isotope As aresult the 120575
18O in precipitation decreases with decreasingair temperature increasing elevation increasing latitudeincreasing distance of vapor transport through the atmo-sphere and increasing precipitation amounts
Figure 4 Stable isotope (120575D and 12057518O) compositions of river water
The isotopic composition of groundwater ranges in minus95simminus67permil (in 120575
18O) and minus682simndash453permil (in 120575D) The equationbetween 120575
18O and 120575D (120575D = 82812057518O + 1074 1198772 = 096119899 = 31) (Figure 5(b)) It is fairly close to that of riverwater The value of 12057518O ranges in minus147simndash123permil and 120575D inminus1050simndash851permil in glacial meltwater The equation between12057518O and 120575D (120575D = 75512057518O + 734 119877
2= 096 119899 =
19) (Figure 5(c)) The stable isotope ratios of hydrogen andoxygen of water samples can provide essential informationabout water dynamics within a given watershed In generalthis is from isotope fractionation by evaporation altitudeeffects and different water sources they received [23] Theslope and the intercept of LMWLwere slightly lower showingdrier and stronger local evaporation conditions Evaporationcaused a differential increase in the 120575D and 120575
18O values ofthe remaining water resulting in a lower slope for the linearrelationship between 120575D and 120575
18O values [41]
43 Temporal Variance of Clminus It can be assumed that mixingprocesses in the catchment determine the isotopic concentra-tion of total runoffHowever the hydrochemical compositionof water is essentially changed as a result of interactions withorganic and inorganicmaterial during its passage through theunsaturated and saturated zones The concentration of Clminusin river water fluctuates in 94sim134mgsdotLminus1 with an averageof 112mgsdotLminus1 The variance of Clminus concentration has muchrelationship with runoff (Figure 6) In spring the springflood caused by snowmelt water makes the soil chemicalions into the river so the concentration of Clminus is relativelyhigh Afterwards with the increase of snowmelt water glaciermeltwater and precipitation the runoff has been showing adifferent amplitudes increaseWith the increase of runoff thedilution effect of ions has also increased so the concentrationof Clminus decreased Although groundwater recharged thedilution effect outweighs the supply effect During Augustand September the runoff has a considerable decrease
Figure 5 Stable isotope (120575D and 12057518O) compositions of precipitation (a) groundwater (b) and glacial meltwater (c)
the concentration of Clminus decreased because of the weakeningof dilution effect
44 Identification of End Members Applying the method ofgeochemical and isotopic tracing in this paper the runoffcharacteristics and hydrological law at the Gahe station areinvestigated at different periods in 2009 We conducted aperformance of principal component analysis (PCA) on theconcentration data of chloride ion The results showed thatthe chemical tracer exhibits conservative behavior whereasisotopes are geographical source tracers and only changecomposition due to slow fractionation processes [42] 18Obelongs to the group of stable environmental isotopes occur-ring naturally in water It has been widely used to separate
storm flow into proportions of event and preevent water [1]As part of the water molecule 18O behaves conservativelythat is the combination of chemical and isotopic tracersallows identifying the origin of water pathways
For the three-component hydrograph separation thechoice of a suitable tracer constellation to explain thechemical changes in discharge during a storm as well asto determine and to identify dominant sources flow pathsand residence times in the catchment becomes increas-ingly important The study suggests that hydrological tracerchlorine concentration and 18O can be used under certainhydrological and lithological conditions A system of alge-braic equations is introduced that enables a three-componenthydrograph separation by using 18O and chlorine These
Advances in Meteorology 7
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
Figure 5 Stable isotope (120575D and 12057518O) compositions of precipitation (a) groundwater (b) and glacial meltwater (c)
the concentration of Clminus decreased because of the weakeningof dilution effect
44 Identification of End Members Applying the method ofgeochemical and isotopic tracing in this paper the runoffcharacteristics and hydrological law at the Gahe station areinvestigated at different periods in 2009 We conducted aperformance of principal component analysis (PCA) on theconcentration data of chloride ion The results showed thatthe chemical tracer exhibits conservative behavior whereasisotopes are geographical source tracers and only changecomposition due to slow fractionation processes [42] 18Obelongs to the group of stable environmental isotopes occur-ring naturally in water It has been widely used to separate
storm flow into proportions of event and preevent water [1]As part of the water molecule 18O behaves conservativelythat is the combination of chemical and isotopic tracersallows identifying the origin of water pathways
For the three-component hydrograph separation thechoice of a suitable tracer constellation to explain thechemical changes in discharge during a storm as well asto determine and to identify dominant sources flow pathsand residence times in the catchment becomes increas-ingly important The study suggests that hydrological tracerchlorine concentration and 18O can be used under certainhydrological and lithological conditions A system of alge-braic equations is introduced that enables a three-componenthydrograph separation by using 18O and chlorine These
Advances in Meteorology 7
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
Table 1 Mean maximum minimum and standard deviation values for the concentration of isotope and chloride ion in river waterprecipitation groundwater and glacial meltwater
Mean Max Min SD Mean Max Min SD Mean Max Min SDRiver water minus92 minus85 minus99 036 minus627 minus582 minus689 349 112 134 94 117Precipitation minus102 minus83 minus130 156 minus697 minus541 minus951 125 05 08 02 022Ground water minus81 minus67 minus95 076 minus564 minus453 minus682 644 163 197 110 220Glacial meltwater minus132 minus123 minus147 073 minus921 minus851 minus1050 563 22 32 12 056
0
20
40
60
80
100
120
140
160
90
95
100
105
110
115
120
125
130
135
4 5 6 7 8 9Month
Runoff
Runo
ff (m
3s
)
Concentration of Clminus
Clminus
(mg
L)
Figure 6 Monthly runoff and concentration of Clminus (with errorbars)
alternative tracers should however be verified against moreconventional tracers before use as the behavior depends onspecific characteristics of solutes
The basic assumption in EMMA is that the stream wateris a discretemixture of its sourcesThe sourcesmust thereforebe of sufficiently different concentrations compared withthe stream water We projected the average values of tracerchloride ion and 18O of three end members in triangle totest and verify the independence (Figure 7) It shows thatmost of streamwater observations fall into the triangle that isspanned by three end members (precipitation groundwaterand glacial meltwater) However there exist some streamwater observations that lie outside of the triangle In manyother studies that apply EMMA such a situation has beendescribed [14 22ndash24 35]These outliers result from a numberof factors including (1) uncertainty in field sampling andlaboratory analyses (2) lack of temporal invariance of endmembers or (3) the expression of different end memberin the mixture as water source areas change temporallyOverall the result can lead to over- or underprediction of thecontributions of each end member to the stream water andshould be understood as a source of uncertainty
45 Contribution of End Members to Runoff Owing tothe geological and geomorphological genesis of the study
minus135
minus125
minus115
minus105
minus95
minus85
minus75
minus65
00 30 60 90 120 150 180
PrecipitationGroundwater
Glacial meltwaterRiver water
Clminus (mgL)
12057518
O (permil
)
Figure 7 Stream water observations and the average values oftracers Clminus and 18O of three endmembers that spanned the triangle
site there are at least three runoff sources having distincthydrological characteristics Results obtained by the use ofthe three-component mixing model are shown in Figure 8Based on the concentration data of isotope and chloride ion(Table 1) the contributions of each end member to riverwater were calculated according to the steady-state massbalance equations of water and tracer fluxes (equations (2))Isotopic hydrograph separation shows that the contributionof groundwater precipitation and glacial meltwater is 667199 and 134 respectively The study indicated thatgroundwater dominated runoff in the headwater area of ShuleRiver Basin And the roles of glacier meltwater should besignificantly noticed in water resource management in thiscatchmentThe glaciers are the headwaters ofmany rivers andthey affect the water discharge of large rivers [43]
Despite the reasonable illustration of the qualitativebehavior of runoff components an exact quantification ofrunoff components contributing remained difficult and isstrongly related to the determination of tracer concentrations
8 Advances in Meteorology
0
10
20
30
40
50
60
70
50
100
150
200
250
300
350
400
450
500
Prec
ipita
tion
(mm
)
PrecipitationGlacial meltwaterGroundwater
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
0
91 101
111
121
131
141
151
161
171
181
191
201
211
221
231
241
251
261
271
Julian day
Runo
ff (m
3s
)
Figure 8 The conclusion of hydrograph separation
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
in both runoff sources Owing to the presence of variousuncertainties only qualitative results were achieved Furtherexperimental investigations are needed to define the tracerconcentrations and their variability with greater accuracyFor hydrograph separations in larger scale basins an exten-sive consideration of the spatial variability along with thesuperposition of spatially distributed runoff components is achallenging task for future research Our results prompt us tofocus future work on understanding interannual changes inend member contribution especially in semiarid regions
5 Discussion and Conclusions
The isotopic and chemical values originate from measure-ment methods field data or the expert knowledge of theinvestigators This is reasonable even if the implications ofthe effects are the same since all of them cause an uncertainestimation of the end member concentrations and thus ofthe contribution of different runoff components So-calledend member concentrations need to be defined for everytracer of a specific runoff component for each separationtime step in order to calculate the component proportionsusing mass balance equations for the tracers and the waterHowever the determination of these concentrations is oftenproblematic as it has been shown that they may exhibit hightemporal and spatial variability and always include errorscaused by the analysis [18 21] Therefore the uncertaintyof hydrograph separation results must be addressed Ingeneral large relative uncertainties must be considered whileperforming hydrograph separations Predictive uncertaintyis the primary impediment to progress in this area butcontinued progress is being made to more fully quantifyuncertainty and more fully explore its implications Theimportance of reducing errors that have the largest impacton uncertainty is clearly demonstrated Therefore future
investigations are needed to define with greater accuracyend member tracer concentrations and their spatial andtemporal variability Moving towards an understanding ofuncertainties complexity is a challenging and important taskfor future research in catchment hydrology
Assumed values of the uncertainty in isotopic compo-sition were 119908
119862119904= 02permil 119908
119862119901= 119908119862119890
= 04permil [43] Wecalculated the uncertainty of tracers itself to be 9 Analysessuggested that the uncertainty in the measurement methodwas less important than that in the temporal and spatialvariations of tracer concentrations The uncertainty termsfor precipitation were generally higher than 80 of the totaluncertainty indicating that the 120575
18O values of precipitationaccount for the majority of uncertainty The uncertainty wassensitive when the difference between mixing componentswas small Therefore the variation of tracers and the dif-ference of mixing components should be considered whenhydrograph separation was applied in the basin
Hydrograph separation shows that the contribution ofgroundwater precipitation and glacial meltwater is 667 plusmn
602 199 plusmn 179 and 134 plusmn 120 respectively Thereare 347 glaciers and the area of glaciers is 2945 km2which accounts for 072 of the headwater area Underthe background of global warming rising temperature leadsto the increase of snowmelt and accelerating the retreat ofglaciers which will have a significant impact on regionalrunoff The roles of glacier meltwater should be significantlynoticed in water resource management in this catchment Inaddition to temporal variability a superposition of spatial andtemporal distributed runoff components needs to be consid-ered Moving towards an understanding of this complexityis a challenging and important task for future research incatchment hydrology
Several studies compared the results of two- and three-component separation Wenninger et al (2004) showed adifference of 10 in preevent water contributions betweenthe two methods because the three-component separationaccounted for snow and rain inputs together while the two-component separation accounted for rain inputs only [44]Dense temporal sampling of hydrographs is often challeng-ing especially at remote locations Therefore studies thatinvestigate runoff generation in alpine catchment are rarePionke et al showed that for a 74 km2 watershed three offour monitored storms were dominated by preevent water(55ndash94 in total) [45] DeWalle showed that in a smallercatchment (0198 km2) storm runoff was also dominated bypreevent water contributions 90 over the course of thehydrograph [46] For a 45 km2 catchment Buda found morethan 80 preevent water contributions to the channel stormflow with 67 during the peak flow in a 6 ha catchment [47]A similar range of preevent water contributions (80 60)was reported by Munyanzea et al (2012) for two mesoscalecatchments (1293 km2 2574 km2) in Rwanda [48] Dong etal have used the recursive digital filtermethod and smoothedminimummethod to separate base flow based on daily runoffdata from 1954 to 2009 in the upper reaches of the Shule RiverBasin the recursive digital filter and smoothed minimummethod were used for base flow separation The base flow
Advances in Meteorology 9
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
index is different between the calculation results from the twomethods (077 and 066) [49]
One shortcoming of this study is that the identificationof end members is limited to data collected during thevegetation period which comprises only 6 months of theyear Moreover we only have the isotope data and chemicalparameters of one hydrologic section so we cannot analyzethe spatial variability The importance of reducing errors thathave the largest impact is clearly demonstrated thereforea targeted sampling strategy is required In order to fullycharacterize the range of climatic variability our resultsemphasize the need of continued development of the long-term measurement Our results prompt us to focus futurework on understanding interannual changes in end membercontribution especially in semiarid regions
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This research was financially supported by the National Nat-ural Science Foundation of China (nos 41130638 4127108541401039) The authors are also grateful to all participants inthe field (J H Yang HWu) and in the laboratory (R Xu) fortheir contributionwhich allows this study to progress in goodconditions
References
[1] J M Buttle ldquoIsotope hydrograph separations and rapid deliveryof pre-event water from drainage basinsrdquo Progress in PhysicalGeography vol 18 no 1 pp 16ndash41 1994
[2] M Bonell ldquoSelected challenges in runoff generation researchin forests from the hillslope to headwater drainage basin scalerdquoJournal of the AmericanWater Resources Association vol 34 no4 pp 765ndash785 1998
[3] S Hoeg S Uhlenbrook and Ch Leibundgut ldquoHydrographseparation in a mountainous catchmentmdashcombining hydro-chemical and isotopic tracersrdquo Hydrological Processes vol 14no 7 pp 1199ndash1216 2000
[4] Z S Wang C F Zhou B H Guan et al ldquoThe headwater lossof the western plateau exacerbates Chinarsquos long thirstrdquo AMBIOvol 35 no 5 pp 271ndash272 2006
[5] B G Mark ldquoHot ice glaciers in the tropics are making thepressrdquo Hydrological Processes vol 16 no 16 pp 3297ndash33022002
[6] F K Barthold J Wu K B Vache K Schneider H-G Fredeand L Breuer ldquoIdentification of geographic runoff sources in adata sparse region hydrological processes and the limitations oftracer-based approachesrdquoHydrological Processes vol 24 no 16pp 2313ndash2327 2010
[7] S Uhlenbrook and S Hoeg ldquoQuantifying uncertainties intracer-based hydrograph separations a case study for two-three- and five-component hydrograph separations in a moun-tainous catchmentrdquoHydrological Processes vol 17 no 2 pp 431ndash453 2003
[8] J Cable K Ogle and D Williams ldquoContribution of glaciermeltwater to streamflow in the Wind River Range Wyominginferred via a Bayesian mixing model applied to isotopicmeasurementsrdquoHydrological Processes vol 25 no 14 pp 2228ndash2236 2011
[9] Y Kong and Z Pang ldquoEvaluating the sensitivity of glacier riversto climate change based onhydrograph separation of dischargerdquoJournal of Hydrology vol 434-435 pp 121ndash129 2012
[10] T Pu Y He G Zhu N Zhang J Du and C Wang ldquoCharac-teristics of water stable isotopes and hydrograph separation inBaishui catchment during the wet season in MtYulong regionsouth western ChinardquoHydrological Processes vol 27 no 25 pp3641ndash3648 2013
[11] C Kendall and T B Coplen ldquoDistribution of oxygen-18 anddeuteriun in river waters across the United StatesrdquoHydrologicalProcesses vol 15 no 7 pp 1363ndash1393 2001
[12] F J Liu M Williams and N Caine ldquoSources waters and flowpaths in an alpine catchment Colorado Front Range USArdquoWater Resources Research vol 40 Article IDW09401 2004
[13] J J Gibson T W D Edwards S J Birks et al ldquoProgress inisotope tracer hydrology in CanadardquoHydrological Processes vol19 no 1 pp 303ndash327 2005
[14] C Wels R J Cornett and B D Lazerte ldquoHydrograph separa-tion a comparison of geochemical and isotopic tracersrdquo Journalof Hydrology vol 122 no 1ndash4 pp 253ndash274 1991
[15] P Durand M Neal and C Neal ldquoVariations in stable oxygenisotope and solute concentrations in small sub-mediterraneanmontane streamsrdquo Journal of Hydrology vol 144 no 1ndash4 pp283ndash290 1993
[16] N C Christophersen C Neal R P Hooper R D Vogt andS Andersen ldquoModelling streamwater chemistry as a mixtureof soilwater end- membersmdasha step towards second-generationacidification modelsrdquo Journal of Hydrology vol 116 no 1ndash4 pp307ndash320 1990
[17] G F Pinder and J F Jones ldquoDetermination of the ground-water component of peak discharge from the chemistry of totalrunoffrdquo Water Resources Research vol 5 no 2 pp 438ndash4451969
[18] J J McDonnell M Bonell M K Stewart and A J PearceldquoDeuterium variations in storm rainfall implications for streamhydrograph separationrdquo Water Resources Research vol 26 no3 pp 455ndash458 1990
[19] W G Mook ldquoEnvironmental isotopes in the hydrologicalcycle principles and applications Volume III surface waterInternational Hydrological Programmerdquo Technical Documentsin Hydrology 39 IAEA Vienna Austria 2001
[20] P K Aggarwal ldquoIsotope hydrology at the International AtomicEnergy AgencyrdquoHydrological Processes vol 16 no 11 pp 2257ndash2259 2002
[21] N Christophersen and R P Hooper ldquoMultivariate analysis ofstream water chemical data the use of principal componentsanalysis for the end-member mixing problemrdquoWater ResourcesResearch vol 28 no 1 pp 99ndash107 1992
[22] F J Liu R Parmenter P D Brooks M H Conklin and R CBales ldquoSeasonal and interannual variation of streamflow path-ways and biogeochemical implications in semi-arid forestedcatchments in Valles Caldera New Mexicordquo Ecohydrology vol1 no 3 pp 239ndash252 2008
[23] J Chaves C Neill S Germer S G Neto A Krusche and HElsenbeer ldquoLand management impacts on runoff sources insmall Amazon watershedsrdquo Hydrological Processes vol 22 no12 pp 1766ndash1775 2008
10 Advances in Meteorology
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
[24] V Acuna and C N Dahm ldquoImpact of monsoonal rains onspatial scaling patterns in water chemistry of a semiarid rivernetworkrdquo Journal of Geophysical Research G Biogeosciences vol112 no 4 Article ID G04009 2007
[25] B Ladouche A Probst D Viville et al ldquoHydrograph separationusing isotopic chemical and hydrological approaches (Streng-bach catchment France)rdquo Journal ofHydrology vol 242 no 3-4pp 255ndash274 2001
[26] D Yang C Li B Ye et al Chinese Perspectives on PUB andthe Working Group Initiative Predictions in Ungauged BasinsInternational Perspectives on the State of the Art and PathwaysForward vol 301 IAHS Publication 2005
[27] J Klaus and J J McDonnell ldquoHydrograph separation usingstable isotopes review and evaluationrdquo Journal of Hydrologyvol 505 pp 47ndash64 2013
[28] C H Leibundgut ldquoTracer-based assessment of vulnerability inmountainous headwatersrdquo in Hydrology Water Resources andEcology in Headwaters vol 248 pp 317ndash326 1998
[29] Z Kattan ldquoCharacterization of surface water and groundwaterin the Damascus Ghotta basin hydrochemical and environ-mental isotopes approachesrdquoEnvironmental Geology vol 51 no2 pp 173ndash201 2006
[30] W Liu S Chen X Qin et al ldquoStorage patterns and control ofsoil organic carbon and nitrogen in the northeastern margin ofthe QinghaindashTibetan Plateaurdquo Environmental Research Lettersvol 7 no 3 Article ID 035401 2012
[31] M J Liu TDHan JWang et al ldquoVariations of the componentsof radiation in permafrost region of the upstream of ShuleRiverrdquo Plateau Meteorology vol 32 no 2 pp 411ndash422 2013
[32] Y Sheng J Li J C Wu B S Ye and J Wang ldquoDistributionpatterns of permafrost in the upper area of Shule River withthe application of GIS techniquerdquo Journal of China Universityof Mining and Technology vol 39 no 1 pp 32ndash40 2010
[33] X Xie G J Yang Z-RWang and J Wang ldquoLandscape patternchange in mountainous areas along an altitude gradient in theupper reaches of Shule Riverrdquo Chinese Journal of Ecology vol29 no 7 pp 1420ndash1426 2010
[34] M G Sklash and R N Farvolden ldquoThe role of groundwater instorm runoffrdquo Journal of Hydrology vol 43 no 1ndash4 pp 45ndash651979
[35] M J Hinton S L Schiff and M C English ldquoExaminingthe contributions of glacial till water to storm runoff usingtwo- and three-component hydrograph separationsrdquo WaterResources Research vol 30 no 4 pp 983ndash993 1994
[36] A RodheThe origin of streamwater traced by oxygen-18 [PhDthesis] Uppsala University 1987 UNGI Report Series A No 41
[37] D Genereux ldquoQuantifying uncertainty in tracer-based hydro-graph separationsrdquoWater Resources Research vol 34 no 4 pp915ndash919 1998
[38] A L James and N T Roulet ldquoInvestigating the applicabilityof end-member mixing analysis (EMMA) across scale a studyof eight small nested catchments in a temperate forestedwatershedrdquo Water Resources Research vol 42 no 8 Article IDW08434 2006
[39] W Dansgaard ldquoStable isotopes in precipitationrdquo Tellus vol 16no 4 pp 436ndash468 1964
[40] J Jouzel K Froehlich and U Schotterer ldquoDeuterium andoxygen-18 in present-day precipitation data and modellingrdquoHydrological Sciences Journal vol 42 no 5 pp 747ndash763 1997
[41] J Wu Y Ding B Ye Q Yang X Zhang and J Wang ldquoSpatio-temporal variation of stable isotopes in precipitation in the
Heihe River Basin Northwestern Chinardquo Environmental EarthSciences vol 61 no 6 pp 1123ndash1134 2010
[42] J K Wu Y J Ding J H Yang et al ldquoStable isotopes in differentwaters during melt season in Laohugou Glacial CatchmentShule River basin Northwestern Chinardquo Journal of Mountain-ous Science In press
[43] T D Yao L Thompson W Yang et al ldquoDifferent glacierstatus with atmospheric circulations in Tibetan Plateau andsurroundingsrdquoNature Climate Change vol 2 pp 663ndash667 2012
[44] J Wenninger S Uhlenbrook N Tilch and C LeibundgutldquoExperimental evidence of fast groundwater responses in a hill-slopefloodplain area in the Black Forest Mountains GermanyrdquoHydrological Processes vol 18 no 17 pp 3305ndash3322 2004
[45] H B Pionke W J Gburek and G J Folmar ldquoQuantifyingstormflow components in a Pennsylvania watershed when 18Oinput and storm conditions varyrdquo Journal of Hydrology vol 148no 1ndash4 pp 169ndash187 1993
[46] D R Dewalle and B R Swistock ldquoDifferences in oxygen-18content of throughfall and rainfall in hardwood and coniferousforestsrdquo Hydrological Processes vol 8 no 1 pp 75ndash82 1994
[47] A R Buda and D R DeWalle ldquoDynamics of stream nitratesources and flow pathways during stormflows on urban forestand agricultural watersheds in central Pennsylvania USArdquoHydrological Processes vol 23 no 23 pp 3292ndash3305 2009
[48] O Munyaneza J Wenninger and S Uhlenbrook ldquoIdentifi-cation of runoff generation processes using hydrometric andtracer methods in a meso-scale catchment in RwandardquoHydrol-ogy and Earth System Sciences vol 16 no 7 pp 1991ndash2004 2012
[49] W W Dong Y J Ding and X Wei ldquoVariation of the base flowand its causes in the upper reaches of the Shule River in theQilian Mountainsrdquo Journal of Glaciology and Geocryology vol36 no 3 pp 661ndash669 2014
