Spatial distribution of oxygen-18 and deuterium in stream waters across the Japanese archipelago

12 Pages·2015·3.13 MB·English

Hydrol.EarthSyst.Sci.,19,1577–1588,2015
www.hydrol-earth-syst-sci.net/19/1577/2015/
doi:10.5194/hess-19-1577-2015
©Author(s)2015.CCAttribution3.0License.
Spatial distribution of oxygen-18 and deuterium in stream waters
across the Japanese archipelago
M.Katsuyama1,2,T.Yoshioka3,andE.Konohira4
1CenterforthePromotionofInterdisciplinaryEducationandResearch,KyotoUniv.,KyotoUniversityHigashiIchijokan,1
YoshidaNakaadachi,Sakyo,Kyoto606-8306,Japan
2GraduateSchoolofAgriculture,KyotoUniv.,KitashirakawaOiwake,Sakyo,Kyoto606-8502,Japan
3FieldScienceEducationandResearchCenter,KyotoUniv.,KitashirakawaOiwake,Sakyo,Kyoto606-8502,Japan
4DLDInc.,Ina,2435KamiYamada,Takatoh,Ina,Nagano396-0217,Japan
Correspondenceto:M.Katsuyama([email protected])
Received:30August2014–PublishedinHydrol.EarthSyst.Sci.Discuss.:30September2014
Revised:12March2015–Accepted:13March2015–Published:31March2015
Abstract. The spatial distribution of oxygen and hydrogen 1 Introduction
isotopiccomposition(δ18Oandδ2H)ofstreamwatersacross
Japanwasclarifiedwithadatasetbycompilingsampledata
obtainedfrom1278forestcatchmentsduringthesummerof The importance of isoscapes, that is, the mapping of large-
2003. Both δ18O and δ2H values showed positive correla- scale spatiotemporal distributions of stable isotope compo-
tionswiththemeanannualairtemperatureandannualevap- sitions in various environments (West et al., 2010), is be-
otranspiration,andnegativecorrelationswithlatitudeandel- ing recognized as providing a framework for fundamental
evation.Deuteriumexcess(d-excess)valuesinstreamwaters and applied research questions in a wide range of fields at
were higher on the Sea of Japan side, and lower on the Pa- large scales. The Global Network for Isotopes in Precipita-
cific Ocean side, of the Japanese archipelago. The d-excess tion(GNIP)databasehasbeenapplied,forexample,tomon-
in precipitation was generally higher in winter and lower in itor climate-change impacts on the character and intensity
summerinJapan.TheSeaofJapansideexperiencesagreat of precipitation (Aggarwal et al., 2012) and to build glob-
dealofsnowfall,andseasonalchangesinmonthlyprecipita- allypredictiveGIS-basedmodelsforprecipitationisoscapes
tion are rather small. In contrast, the Pacific Ocean side ex- (e.g.,www.waterisotopes.org).
periencesalargeamountofrainfallduringsummerwithlow Wassenaar et al. (2009) pointed out, however, that the
levelsofprecipitationduringthewinter.Therefore,thelower GNIPstationsareoftenspatiallydeficientformanyregions
d-excess in stream waters on the Pacific Ocean side reflects thatareofinteresttohydrologistsaswellasecologists.For
summer precipitation, and the higher values on the Sea of example,Wassenaaretal.(2009)mentionedthatMexicohas
Japansideareaffectedbydelayedrechargefromsnowmelt. onlytwoGNIPstations.Moreover,Japanhadalsoonlytwo
Theisoscapesofstreamwaterconnotenotonlyspatiallyin- stations and, unfortunately, both stations in Japan were al-
tegratedbutalsotemporallyintegratedisotopesignalsofpre- ready closed. In addition, long-term monitoring of precipi-
cipitation and provide a framework for addressing applied tation is also required. Consequently, the ground validation
hydrological, ecological, or meteorological research ques- data for these global models are insufficient to compare at
tionsatregionalscales,suchastheeffectsofclimatechange. regional or country-wide scales. Under the circumstances,
Wassenaar et al. (2009) hypothesized that the stable iso-
topic composition of surface water or groundwater, which
integrates longer-term precipitation inputs (Clark and Fritz,
1997),canbeaproxyforprecipitationinfiltrationinput.In-
deed, some research has been undertaken regarding nation-
widesurface-waterandgroundwaterisoscapesandusesthem
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
1578 M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters
asanindicatoroftheprecipitationisoscape(e.g.,theBritish forestontheroadmapbeforethesampling,andverifiedthe
Isles: Darling et al., 2003; the United States: Kendall and inexistenceofartificialpollutionsourcesuchasdams,houses
Coplen, 2001; Finland: Kortelainen and Karhu, 2004; Mex- and/orfarmlands,andproperlychangedthepointsinthefield
ico: Wassenaar et al., 2009). Although Mizota and Kusak- toavoidtheeffectsofthem.Therewerenolakes,swaps,hot
abe(1994)havealreadypresentedthespatialdistributionof springsorotherwaterbodieswhichcouldaffecttheisotope
stable isotope compositions of surface water in Japan, they values in each catchment. The sampling points covered 45
do not discuss the mechanisms underlying the distribution. prefectures;onlytwoprefectures,ChibaandOkinawa,were
In other words, it is insufficient to test the hypothesis re- excluded from the campaign. Samples were collected from
garding the isotope signals of precipitationinput being spa- approximately30catchmentsineachprefectureand,finally,
tially and temporally integrated in the stream water output. 1278 forested headwater catchments were selected (Fig. 1).
Global warming will dramatically change the hydrological Thecatchmentareasrangedfrom0.05to136.8km2andwere
responsesofwatersheds.Thesechangesofthehydrological 5.3km2 onaverage.Intotal,95.3 %ofthecatchmentswere
responses are driven by temperature and precipitation pat- smallerthan15km2.Thesamplingprocedureswereunified
ternsthatwillaffectthetemporalandspatialdistributionsof between all 11 researchers prior to sampling. During the
river source water over time (Marshall and Randhir, 2008). campaign,grabsamplesofstreamwaterwerecollectedonce
Therefore,thelinkagebetweentheprecipitationandsurface ineachcatchment.Asthevaluescanbeaffectedbyprecipita-
water at each point in time should be clarified because sur- tion,wecollectedthesamplesduringbaseflowconditionand
face water is the most important water resource. At finer avoided to sample during and just after the precipitation to
scales,thetemporalvariationinthestableisotopesignalsof unifythecollectionconditionsasmuchaspossiblebetween
precipitation and stream water have been used to estimate sites. The collected samples were immediately filtered and
the mean residence time of stream water within catchments preservedbyfreezinginpolycarbonatebottlesat−10◦ until
(McGuireandMcDonnell,2006;Dunnetal.,2008;Tetzlaff analyzedin2008.
et al., 2011); however, few studies have been conducted in The stream water samples were analyzed for both δ18O
Japan(e.g.,Katsuyamaetal.,2010).Theresultsofthesees- andδ2HbytheColoradoPlateauStableIsotopeLaboratory
timatesmaychangeduetofuturechangesinthehydrological using an Off-Axis Integrated Cavity Output Spectroscopy
responses of the watershed. Therefore, the establishment of liquid water isotope analyzer (Los Gatos Model 908-0008).
a nation-wide and spatially dense stream water isotope net- Themeasurementprecision(standarddeviation)was±0.2%
workforJapan,whichhasawiderangeofclimaticandge- and±0.8%forδ18Oandδ2H,respectively.
ographicalconditionsoverasmallarea,mayprovidespatial
isotope information fundamental for the application of iso- 2.2 Isotopedataonprecipitationandclimate
topesinhydrologicalstudies. conditions
Herewepresentthestreamwaterδ18Oandδ2Hisoscapes
Wecollectedisotopedataonprecipitationfromthepublished
oftheJapanesearchipelagoandprovidemultivariateregres-
literature and unpublished data kindly offered by many re-
sionanalysesusingkeyenvironmentalandgeographicalpa-
searchers,inadditiontoouroriginaldata.Thepolicyforcol-
rameterstodeterminewhichvariablesarethekeydriversof
lecting data was that both δ18O and δ2H were to be mea-
streamwaterisotopicpatterns.Theidentificationofkeypa-
sured monthly or more frequently for over 1 year to calcu-
rameters is essential in evaluating the vulnerability of hy-
late the mean annual weighted value of the successive pre-
drological responses in the watershed to climate change.
cipitation inputs. The data were collected from 14 prefec-
Moreover,bycomparingthedatawithexistingprecipitation-
tures(Fig.1andTable1).Inthreeoftheseprefectures,Shiga
isotopedata,weconsidertheadvantageofusingstreamwa-
(no. 6 in Fig. 1), Nara (8), and Tottori (9), the precipitation
ter isoscapes as an integrated indicator of precipitation for
samplinghasbeencontinuous.InShiga,themonitoringbe-
futureisotopichydrologystudies.
ganin1997attheKiryuExperimentalWatershed(Kabeyaet
al.,2007;Katsuyamaetal.,2010).Samplingbeganin2004at
the Mt. Gomadan Experimental Forest in Nara (Katsuyama
2 Methods
et al., 2008; Fukushima and Tokuchi, 2009), and in 2011 at
theHiruzenExperimentalForestofTottoriUniversity(Haga
2.1 Streamwatersamplingandmeasurement
andKatsuyama,unpublisheddata).TheHiruzenForestislo-
cated on the Okayama side of the Okayama–Tottori prefec-
The sampling campaign Japan-Wide Stream Monitoring
turalborderandthesampleswerecollectedfromthemoun-
(JWSM)2003(Konohiraetal.,2006)wasconductedduring
tain peak. Therefore, we term this station Tottori to clearly
the summer, from 1 July to 11 October of 2003, by 11 re-
distinguishitfromOkayama(10).Atthesethreestations,the
searchers.Allsampleswerecollectedfromforestedheadwa-
corresponding stream water sampling has also been contin-
ter streams to avoid the influence of anthropogenic impacts
uous at the outlet of each catchment. The catchment area is
suchasagricultureorurbaneffects.Weselectedthepotential
5.99haforShiga,3.15haforNara,and5.9haforTottori.
samplingpointswherethewholecatchmentwascoveredby
Hydrol.EarthSyst.Sci.,19,1577–1588,2015 www.hydrol-earth-syst-sci.net/19/1577/2015/
M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters 1579
Figure1.Indexmapshowinglocationsofthe1278streamwatersamplingpointsandthe14samplingstationsofprecipitation.Lightblue
circles:streamwatersamplingpoints.Blackcircles:precipitationsamplingstation.ThestationnumberscorrespondtothoseinTable1.
The collected samples from five of these prefectures, matedusingthePriestley–Taylormethod(PriestleyandTay-
Shiga (6), Kyoto (7), Nara (8), Tottori (9), and Kochi (12), lor,1972).
weresealedinsmallglassvialsandmaintainedatroomtem-
perature.Amassspectrometer(ThermoElectron,MAT252)
attheCenterforEcologicalResearch,KyotoUniversity,with 3 Results
the CO –H O and H –H O equilibrium methods was used
2 2 2 2
forbothδ18Oandδ2Hanalysis.Themeasurementprecision 3.1 Relationshipbetweenδ18Oandδ2Hvaluesin
(standard deviation) was ±0.05 and ±0.9‰ for δ18O and streamwater
δ2H,respectively.
Themeasuredδ18Oandδ2Hvaluesinstreamwatersamples
Theisotopevaluesofbothstreamwaterandofprecipita-
ranged from −13.7 to −5.9‰ (mean =−9.2‰) and from
tionarereportedaspermil(‰)unitsrelativetotheVienna
−92.2to−35.1‰(mean=−56.6‰),respectively(Table2).
StandardMeanOceanWater.
Aclearlinearrelationshipexistedbetweenδ18Oandδ2Hval-
ues(Fig.2)as
2.3 Geographicalandenvironmentalparameters
δ2H=6.85δ18O+6.11(n=1278,r2=0.89,p<0.001). (1)
The catchment area and elevation of the site were deter-
mined using a 250m digital elevation model for each sam- The data set forms a flattened ellipse around the regression
plingpoint.Meanannualprecipitation(MAP)andmeanan- line.Thisrelationshipissimilartotheresultsfromprevious
nualtemperature(MAT)wereextractedfromMeshclimatic studies in Japan. Machida and Kondo (2003) collected iso-
dataaveragedfortheyearsfrom1971to2000(JapanMete- tope data for 1067 rivers and shallow groundwater sources
orologicalAgency,2002).Thespatialresolutionofthemesh frommanypapersanddatabasesas
climatic data is 1km. Actual evapotranspiration (AET) was
extracted from Ahn and Tateishi (1994), which they esti- δ2H=6.72δ18O+3.94 (n=1067,r2=0.91). (2)
www.hydrol-earth-syst-sci.net/19/1577/2015/ Hydrol.EarthSyst.Sci.,19,1577–1588,2015
1580 M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters
dcba T count
Thed-excessofprecipitationareextractedbyusfromtheoriginalpapers.MeteorologicaldataarefromthenearestobservationstationbytheJapanMeteorologicalAgency(AMeDAS).Thed-excessvaluesforprecipitationareannualweightedmeanswiththeamountofprecipitationfor1year,andJandPmeanSeaofJapansideandPacificOceanside,respectively,andMATmeansmeanannualtemperature. 14KagoshimaPTakakumaExp.Forest,KagoshimaUniv.14.631.513FukuokaJOchozuExperimentalWatershed16.233.612KochiPMt.Takatori13.133.311KagawaPTakamatsuCity16.334.310OkayamaJOkayamaUniv.16.234.79TottoriJHiruzenExp.Forest,TottoriUniv.10.535.38NaraPMt.GomadanExp.Forest8.934.17KyotoJKyotoUniv.15.135.06ShigaJKiryuExp.Watershed13.534.95ToyamaJToyamaCity14.136.74KanagawaPOoborawatersheds,TanzawaMountains15.135.53SaitamaPRisshoUniv.15.036.12IbarakiPMt.Tsukuba9.736.21AkitaJAkitaUniv.11.739.7 No.SideSiteMATLatitud ofprecipitationaStationPrefectureSamplinglocation able1.Comparisonofd-excessvaluesbetweenprecipitationandstreamw Fquige-un1-------cr09876543ey000000002d.isRtreilbauttiioonnsshiipnbstertewameenwδa1te8rO. andδ1δ82OH(‰va)luesandth433221150e05050500ir0000000fre-
thevaluesforstream 130.8130.5133.0134.0133.9133.6135.6135.8136.0137.2139.2139.4140.1140.1 eNLongitude ater. Maanboder(es1oh9va9ell4ro,)watshgeryournedcwalactuelratperdesetnhteeddabtyaMfoizrotsaurafnadceKwusaatke-r
w E
aterarethe 14.612.913.711.311.818.713.912.713.815.413.313.612.017.0 Precip δ2H=7.03δ18O+7.91 (n=298,r2=0.93). (3)
arithmeticmeansof 14.513.713.812.317.621.713.718.520.319.614.012.913.620.3 itationStream d-excess(‰)b BasKatouemtshcapokltimahnbegpeaspr(loe1odi9pn9ett4soa)itnnahdnotdhstehiMseosafitcnuEhtdeiqydrsca.eexa(p2cnte)doeaKdfneoEddnq(dt.3ho)(o.1(s2)Te0hai0enr3e)nM.iunHimzteoobrwtemaresevadneoirdf-,
alldata water tfhroemdastaampprleesseninteddiifnfetrheensteypearervsioanuds ssteuadsioenssw. Oerne tchoelleocthteedr
in
eachprefecture. AsanoandTateno(unpub.)AsanoandChiwa(unpub.)ShinomiyaandSakai(unpuTaseetal.(1997)Yamamotoetal.(1993)HagaandKatsuyama(unpuKatsuyama(unpub.)Katsuyama(unpub.)Katsuyama(unpub.)Satakeetal.(1984)Oda(unpub.)Yabusaki(2010)Yabusakietal.(2008)Kawarayaetal.(2005) precipitationd-excessDatasourceof hwwpmwδ1araiioe8ntltvchlOediepir,pnmsaritonthseaaddehntuifotδdocew2aepnwHtr,aaovmdomcewuferooyeessnrsteptelrthioersmrseaewsmliloeetieaivfntwdbehttlliahheiintnseeegogrcretesaoniadtnwpemcuirhJarcaeaimslnpsyrgesaeeeynnalasitas.nsrtot.e.fiinomTGlantarhlesiatehntityrciieeoparfcnaloblolyrelmayetncw,,podEsaelurleqweern.cfdaatt(tehc1tedoeer)
b.) b.) The data are grouped into 10 regions and the linear re-
gressions are applied to each region (Table 2). The regional
Jitousonoetal.(2010)Ideetal.(2009)ShinomiyaandYoshinaga Sanoetal.(2014)FukushimaandTokuchi(2 Katsuyamaetal.(2010) Odaetal.(2013) sitedescriptionReferenceof dAth–ailiaogtvadhneniossnadiucoorgCynfahhneuuugatsshgeceeeohddokfrrfufeo2soglr(loivGlowopan)wleeu)srae.eetdaTsghniehtwodhreneaifsrnrtoeethrogeeaflrrcodectawsehsssp.etmitsosTJan.(ahlssFlpeeofaseodlrnorFaeeptsiixaengeasd.mfMir3avopneimflddteoeu,rioatnKhrtlthoeeirenlsreoeckglgeioitipo(cwctFnaasos-)l,
(2008) 009) rreagnigoenso,feδspveacliuaelslyafnrdomweCrehupglookttue,dhoandtahereulpatpiverelryegniaornroowf
cc cd, c cd, Re the data ellipse (Fig. 2). The regression lines vary strongly
ma even within each prefecture in the Kinki and Chugoku re-
rk
s
Hydrol.EarthSyst.Sci.,19,1577–1588,2015 www.hydrol-earth-syst-sci.net/19/1577/2015/
M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters 1581
Table2.Rangeofδ18Oandδ2Hvaluesinstreamwatersamplesandlinerregressionsforeachregion.
Region∗ n δ18O δ2H Slope Intercept r2
max. min. max. min.
A Hokkaido 94 −9.0 −13.7 −54.7 −92.2 6.94 6.79 0.87
B Tohoku 167 −7.6 −12.7 −47.9 −79.9 6.27 1.73 0.84
C Kanto-Koshin 226 −7.1 −13.1 −46.1 −89.2 7.07 4.50 0.93
D Hokuriku 124 −7.8 −13.2 −43.7 −87.9 7.76 19.38 0.90
E Toukai 105 −6.8 −12.4 −38.6 −80.1 6.57 3.20 0.96
F Kinki 175 −6.3 −9.9 −37.7 −62.0 5.15 −6.96 0.62
G Chugoku 117 −7.5 −10.1 −44.6 −56.1 2.96 −25.59 0.40
H Shikoku 111 −6.1 −9.8 −35.1 −63.2 8.02 13.91 0.93
I NorthernKyushu 119 −5.9 −9.0 −38.0 −56.7 6.27 0.31 0.87
J SouthernKyushu 40 −6.4 −8.6 −38.3 −56.3 6.80 4.84 0.73
National 1278 −5.9 −13.7 −35.1 −92.2 6.85 6.11 0.89
∗RegionaldivisionisshowninFig.3.
Figure3.TopographyoftheJapanesearchipelagoandindexofregionaldivision.NotetheregionaldivisionisshowninTable2.Thebase
mapisfrom:http://en.wikipedia.org/wiki/File:Japan_topo_en.jpg.
gions (not shown). The slopes of these lines were relatively side, e.g., Nara (6.5) and Wakayama (6.8). These facts may
smaller at the prefectures of the Sea of Japan side – e.g., imply that there are plural potential regression lines within
Shiga (1.9), Kyoto (4.5), Hyogo (3.5), Tottori (5.2), Shi- the data ellipse at each region as results of the contribution
mane(2.3)–comparedtotheprefecturesofthePacificOcean fromdifferentmoisturesources.Tofindthefinalsolutionof
www.hydrol-earth-syst-sci.net/19/1577/2015/ Hydrol.EarthSyst.Sci.,19,1577–1588,2015
1582 M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters
Table3.Correlationmatrixamongδ18OandparametersLAT,ELV, north.Thelowervaluesareobservedinhighmountainareas
ARA,MAT,MAP,andAETmeaninglatitude,elevation,catchment atthecentralpartsoftheHonshuandHokkaidoislands.The
area,meanannualprecipitation,meanannualtemperature,andac- lowest values were observed at the Mt. Daisetsu region of
tualevapotranspiration,respectively. Hokkaido.Ontheotherhand,thevaluesarehigherinsouth-
westernJapan.
δ18O LAT ELV ARA MAT MAP AET Deuterium excess (d-excess, d=8δ2H-δ18O; Dansgaard,
δ18O 1 1964) is known and provides information about the climate
LAT −0.65 1 conditions of the moisture sources. The d-excess values in
ELV −0.55 −0.06 1 streamwaterwereclearlydividedbythebackbonemountain
ARA −0.31 0.12 0.2 1
MAT 0.88 −0.73 −0.53 −0.35 1 rangesoftheJapanesearchipelago;andthevaluesinstream
MAP 0.32 −0.43 0.01 0.1 0.25 1 waterwereloweronthePacificOceansideandhigheronthe
AET 0.61 −0.88 −0.02 −0.12 0.69 0.44 1 SeaofJapanside(Fig.5).Thed-excessvaluesinstreamwa-
tersamplesrangedfrom0.9to26.9‰(mean=16.6‰).The
specifically depleted d-excess values (d<8) were only ob-
this question, however, we need more detailed observations served in the Gunma Prefecture located at the northern end
ateachprefecturescale.Inotherregions,theslopeswererel- of the Kanto Plain, while the highest values were observed
ativelysimilar(about6–8);however,theinterceptswerevar- intheNiigataPrefectureintheHokurikudistricts,theheav-
ied. These results mean that the ellipse of the data in Fig. 2 iestsnowfallareainJapan.Thispatternofd-excessinsome
iscomposedofmanylocalregressionlinesfortheindividual parts of Japan was reported previously in the pioneer work
regions.KendallandCoplen(2001)alsofoundtheimbricate by Waseda and Nakai (1983). They found that the d-excess
nature of the LMWLs (local meteoric water lines) at states ofsurfacewatersincentralandnortheasternJapantendedto
in the United Sates relative to the GMWL (global meteoric increasecontinuouslyfromthePacificOceansidetotheSea
water line). This fact means that our data is different from ofJapanside,rangingfrom9.1to22.4.
the GMWL but this is not unexpected because the latter is
“comprehensive”and“global”. 3.3 Seasonalvariationofd-excessvaluesin
InJapan,nationwidesystematicobservationsofδ18Oand precipitationandstreamwater
δ2H in precipitation have not been carried out. Under such
circumstances,Taseetal.(1997)presentedalocalmeteoric Figure6aandbshowthelong-termvariationofmonthlyd-
water line for Japan from the observations made at 16 sta- excess in precipitation and stream water observed at Tottori
tionslocatedintheKantoregionandsouthwestJapanas (Sea of Japan side, station no. 9 in Table 1), Shiga (Sea of
Japanside,stationno.6),andNara(PacificOceanside,sta-
δ2H=7.3δ18O+8.6(r2=0.89). (4)
tion no. 8). The panels of Fig. 6c show typical examples of
monthly d-excess values in precipitation and stream water
Comparing the regressions for stream water (Eqs. 1–3) and
with monthly precipitation and air temperature observed at
forprecipitation(Eq.4),theequationsappeargenerallysimi-
Tottori in 2011, Shiga in 2008, and Nara in 2006. The cli-
lar,suggestingthattheisotopiccompositionsofprecipitation
mateconditionsareclearlydifferentamongthesestations.In
arereflectedintheisotopiccompositionsofstreamwaterat
Tottori,muchsnowfallsfromDecembertoMarchwithlow
thenationalscale.Thesmallerslopesandinterceptionsinthe
air temperatures. In Shiga, there is less snowfall but much
formerequationsreflecttheeffectsofevaporationduringin-
morerainfallduringsummer.Summerrainfallismoreplen-
filtrationprocesses.However,asmentionedabove,theellipse
tifulinNara.However,similarsinusoidald-excessvariations
ofdatainFig.2iscomposedofmanylocalregressionlines;
in precipitation were repeated at these three stations every
therefore,moreprecipitationdataareneededtointerpretthe
year, i.e., higher during winter and lower during summer.
meaningofslopesatregionalscalestoallowtheseresultsto
The sinusoidal pattern is caused by the contribution of the
bediscussedindetail.
dualmoisturesourcespredominantlyfromthePacificOcean
3.2 Spatialdistributionofδ18O,δ2H,andd-excess insummerandpredominantlyfromtheSeaofJapaninwin-
valuesinstreamwater ter(WasedaandNakai,1983;Araguás-Araguásetal.,1998).
Moreover, Tase et al. (1997) also reported that this sea-
The Japanese archipelago is elongated from northeast to sonal pattern was commonly observed at six stations in the
southwest, and mountains form its backbone (Fig. 3). High Kanto, Shikoku, and Kyushu regions (see also Fig. 3). Un-
mountains are located at the centers of the Honshu and fortunately,wedidnothavesufficientdatafrom2003when
Hokkaido islands, named the Japan Alps (near 35–37◦N, stream water sampling was conducted. Therefore, we will
136–139◦E)andMt.Daisetsuregion(43◦40(cid:48)N,142◦51(cid:48)E), compare the precipitation values observed in various years
respectively. withthestreamwatervalues,inthefollowing.
Thespatialdistributionsofδ18Oandδ2Hareverysimilar Our observation was conducted from July to October.
(Fig. 4a, b). Generally, the values decreased from south to The isotope signature in stream water should have seasonal
Hydrol.EarthSyst.Sci.,19,1577–1588,2015 www.hydrol-earth-syst-sci.net/19/1577/2015/
M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters 1583
Figure4.Spatialdistributionof(a)δ18O(leftpanel)and(b)δ2H(rightpanel)valuesinstreamwater.
50 Shiga Nara Tottori
40 (a) Precipitation
30
20
10
0
50 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
40 (b) Streamwater
30
20
10
0
Precipitation (mm/month)1234560000000000000 1(c2)34567891(T20o01t1t1o11r)i2 123456789(1S20h01i0g18a1) 2 123456789(12N00a01ra61)12-0511223505050℃Air Temp. () 01234500000（d-excess‰)
Month
Prde-ceixpciteastison ◆ strde-aemxc ewsaster Precipitation ○ Air Temperature
Figure6.Seasonalvariationsofd-excessvaluesinprecipitationand
streamwater.(a)Long-termvariationofmonthlyd-excessinpre-
cipitation.(b)Long-termvariationofmonthlyd-excessinprecipita-
tion.(c)Typicalexamplesofmonthlyd-excessvalueswithmonthly
precipitationandairtemperature.
Figure5.Spatialdistributionofd-excessvaluesinstreamwater.
and0.14inNara.TheCVscalculatedwiththedatafromJuly
toOctoberforprecipitationandstreamwaterwere0.16and
0.07inTottori,0.27and0.03inShiga,and0.48and0.07in
variation and the samples may be biased depending on the Nara.Certainly,wecannotconsidertheseasonalityinstream
date they were taken to some extent. However, as shown in water for all of our sampling; however, these values imply
Fig.6b,theseasonalityinstreamwaterisclearlydampened that the samples are less biased depending on the date they
comparedtothatinprecipitation(Fig.6a)inallstations.The weretakencomparedtotheseasonalityinprecipitation.
coefficientofvariation(CV)calculatedwiththe1-yeardata Thedampingoftheseasonalityinstreamwaterisaresult
(Fig.6c)forprecipitationandstreamwaterineachsitewere ofthehydrologicalprocesseswithinthecatchment.Thesea-
compared. The CV for precipitation and stream water were sonalityofd-excessvaluesissometimesusedtoestimatethe
0.50 and 0.09 in Tottori, 0.43 and 0.07 in Shiga, and 0.62 waterresidence(andtransit)times(Kabeyaetal.,2007;Lee
www.hydrol-earth-syst-sci.net/19/1577/2015/ Hydrol.EarthSyst.Sci.,19,1577–1588,2015
1584 M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters
-5 (a) (b) (c) -5 tendingnorthandsouth.TheAETwascalculatedbyPriest-
-7 -7 ley and Taylor’s (1972) method, which increases in propor-
-9 -9 tion to the net radiation. Thus, AET generally decreases at
higher latitudes. Indeed, the correlation coefficient between
-11 -11
latitudeandAETwasveryhigh(Table3,r=−0.88).There-
--1153 δ18rO2 = = 0 0.7.474,M pA<T0 . 0- 0113.5 δ18Or2 == 00..1000,0 p8<M0A.P00-110.8 δ18Or2 == 00..03076, p7<A0E.T0 0-114.8 --1153 fore,thepositivecorrelationbetweenδ18OandAET(Fig.7c)
0 2 4 6 8 10121416 0 10002000300040005000 0 300 600 900 12001500 covers both the temperature effect (Fig. 7a) and the lati-
Mean Annual Air Temp. (deg.C) Mean Annual Prep. (mm) AET (mm)
-5 (d) δ18O = -0.35LAT + 3.4 (e) δ18O = -0.0028ELV -8.0 (f) δ18O = -0.067ARA - 8.8 -5 tude effect (Fig. 7d). The scatter around latitude 35–37◦N
-7 r2 = 0.43, p<0.001 r2 = 0.30, p<0.001 r2 = 0.10, p<0.001 -7 inFig.7dreflectstheelevationeffectaroundtheJapanAlps
-9 -9 area; and the relationship between δ18O and ELV around
-11 -11 this area (35–37◦N, 136–139◦E) is δ18O=−0.0027 ELV
– 8.4 (r2=0.60, p<0.01; figure not shown). The slope of
-13 -13
theregressionlinefortheelevationeffect(Fig.7e),i.e.,the
-15 -15
303234363840424446 0 500 1000 1500 2000 0 20 40 60 80100120140 isotopic lapse rate of stream water, was −0.28‰100m−1,
Latitude Elevation (m) Catchment Area (km2)
which is the same as the global isotopic lapse rate of pre-
Figure7.Correlationsofδ18Oandenvironmentalorgeographical cipitation (Poage and Chamberlain, 2001). This result sug-
parameters.
geststhatstreamwatersretainthepropertiesofprecipitation
and, thus, the spatial patterns of stream water samples may
be a suitable proxy for precipitation. However, the amount
etal.,2007;KimandJung,2014),andthesmallerseasonality
effect, which is commonly observed to be negatively corre-
instreamwatergenerallymeansalongerresidencetime.The
latedwithprecipitation,wasnotasclear(Fig.7b).Moreover,
control factors of residence times are actively argued in the
noclearrelationshipwasfoundbetweenδ18Oandcatchment
scientific community; for example, the geomorphic factors
area(ARA)(Fig.7f).
(McGuireetal.,2005;Tetzlaffetal.,2009)andthebedrock
Basedontheserelationships,amultipleregressionmodel
permeability(Katsuyamaetal.,2010)canbecontrolfactors.
was developed to identify the controls of environmental
However, it is not necessarily controlled by simple param-
(MAT, MAP, and AET) and geographical (LAT, ELV, and
eters such as recharge area. In other words, the clear spa-
ARA) parameters on stream water δ18O. As noted above,
tialdistributionfoundinFigs.4and5iscorrect,thoughthe
however, AET was highly controlled by latitude and there-
residence time of our samples must be different from each
forewasnotappropriateasapredictorvariable.Theregres-
otherbecauseofthecomplexrelationshipbetweenprecipita-
sionequationthatconsideredtheotherfivedescriptorsisas
tionandstreamwaterineachcatchment.
follows:
δ18O=−0.18LAT−0.0017ELV−0.015ARA+0.21MAT
4 Discussion
+0.00022MAP−4.52(r2 = 0.81,p<0.001). (5)
4.1 Correlationsbetweenδ18Oandenvironmentalor
geographicalparameters LAT,ELV,andMAT,havehighercorrelationswithδ18Othan
theothertwoparameters(Fig.7,Table3).Consideringonly
Generally, the δ18O of precipitation is affected by various thesethreeparameters,theregressionmodelwaschangedto
environmentalandgeographicalvariables.Here,wewilldis-
cusstheeffectontheδ18Oofstreamwaterontheseparam- δ18O=−0.19LAT−0.0018ELV+0.22MAT−3.75
eterstodiscusswhetherthestreamwatercanbetreatedasa (r2= 0.80,p<0.001). (6)
proxy of precipitation. Stream water δ18O had strong posi-
tivecorrelationswithMAT(Fig.7a)andAET(c),aswellas Similarly,theregressionmodelusedforstreamwaterδ2His
a negative correlation with latitude (LAT) (d) and elevation asfollows:
(ELV)(e),variablesthatarecommonlyusedtodescribethe
temperature, latitude, and elevation effects on precipitation δ2H=−0.88LAT−0.017ELV−0.12ARA+1.36MAT
isotopes. These relationships are similarly found in rivers +0.0041MAP−39.94(r2 = 0.84,p<0.001) (7)
acrosstheUnitedStates(KendallandCoplen,2001),andthe
trends with MAT and LAT in groundwater isotopic compo- δ2H=−1.39LAT−0.018ELV+1.22MAT
sitionshavealsobeenobservedinFinland(Kortelainenand −10.89(r2 = 0.79,p<0.001). (8)
Karhu,2004).
Table3showsthecorrelationmatrixamongδ18Oandpa- Theseequationssufficientlyexplaintheobservedstreamwa-
rameters. The correlation between δ18O and MAT is partic- ter δ18O and δ2H. The key drivers of stream water iso-
ularly strong (Fig. 7a). The large variation in MAT reflects topic patterns are, especially, the two geographical param-
the geographical features of the Japanese archipelago ex- eters of LAT and ELV, and one environmental parameter,
Hydrol.EarthSyst.Sci.,19,1577–1588,2015 www.hydrol-earth-syst-sci.net/19/1577/2015/
M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters 1585
MAT.Wassenaaretal.(2009)mentionedthattheregression est(Kondoetal.,1992;Komatsuetal.,2008),andthenum-
modelapproachissuitableforothercountriesandregionsin berwaslargerattheSeaofJapanside(Kondoetal.,1992).
whichGNIPstationsarelacking.Theobservedandpredicted Thus,thecanopyinterceptioncanmakethed-excessvalues
data can be linked to other investigations, such as of eco- lowerattheSeaofJapanside.Ontheotherhand,theevapo-
logical and forensic isotope applications (Wassenaar et al., rationfromforestfloorestimatedinJapanisgenerallysmall
2009;Bowenetal.,2009,2011).Therefore,thestreamwater and negligible. For example, Tsujimura and Tanaka (1998)
isotopiccompositionsinJapanpredictedbytheseequations estimated the value as 2% of annual precipitation in cen-
areapplicabletootherdisciplinesinadditiontohydrological tralJapan.KubotaandTsuboyama(2004)showedthevalues
studies. were below 10% of annual throughfall in both mature and
young forests in Kanto region. These effects of evaporation
4.2 Comparisonofd-excessvaluesinrainwaterand maybereflectedinthesmallerslopesandinterceptionsinthe
streamwater regressionforstreamwater(Eq.1)thanintheregressionfor
precipitation(Eq.4).However,eventheeffectsoftheevapo-
Acomparisonofisotopiccompositionwithprecipitationon ration,especiallyofthecanopyinterception,areconsidered,
a regional scale showed the connection between the source thedifferenceofd-excessvaluesattheSeaofJapansideand
waterofthestreamandthemeteoricinputsignal(Clarkand atthePacificOceansideisclear.Thisfactalsosupportsthat
Fritz,1997;Duttonetal.,2005).Table1showsthecompari- thedifferenceishighlycontrolledbytherechargeprocessin
sonofd-excessvaluesbetweenprecipitationandstreamwa- eachregion.
ter.OnthePacificOceanside,thed-excessvaluesofprecipi- The comparison of d-excess values in precipitation with
tationandstreamwaterwerenearlyequal:thedifferencewas stream water demonstrates that although meteoric water
less than 1‰. On the Kanto Plain (stations 2, 3, and 4), the has relatively spatially homogeneous isotopic compositions
values were very close. On the Sea of Japan side, however, (Taseetal.,1997),therechargedstreamwaterdoesnotnec-
thevaluesofstreamwaterwereclearlyhigherthanthoseof essarily reflect the pattern of the meteoric water. The de-
precipitation, with a difference of more than 3‰, except in layedcontributionofsnowmeltcontrolledtheisotopiccom-
theFukuokaPrefecture.Theseresultsmeanthatseasonalbi- positions of stream water in snowy regions. As there is no
ases exist to recharge on the Sea of Japan side. Compared clear relationship between the catchment areas or elevation
tothePacificOceanside,theSeaofJapansideexperiences with d-excess values (figures not shown), the difference in
a great deal of snowfall, and the d-excess of meteoric input theelevationoftheactualstreamwaterrechargeareacannot
was higher in winter (Fig. 6a, c). Thus, the water recharge be the caseof differences in d-excesssignatures. Moreover,
fromlatesnowmeltmayhaveaffectedsummerstreamwater, themeanresidencetimeofstreamwaterisnotcontrolledby
eventhoughourstreamwatersamplingwasconducteddur- catchment size (e.g., Tetzlaff et al., 2009; Katsuyama et al.,
ingthesummer(fromJulytoOctober).FukuokaPrefecture, 2010);therefore,thedelayedcontributionofsnowmeltisnot
an exception, is located on Kyushu Island and has a warm controlledbygeographicalparameters.Todiscussthismore
climatewithlesssnowfall.Thus,thed-excessoftheprecipi- directly,weneedanationwidesystematicdatasetofisotopic
tationandofthestreamwaterwassimilar.Therefore,inpar- compositionsinprecipitation.Taseetal.(1997)reportedthe
ticular on the Sea of Japan side, we must take into account only example of such observation. However, unfortunately,
thewintersnowpackandspringsnowmeltwhenconsidering theirstationsweremainlylocatedonthePacificOceanside,
the recharge processes of the stream. The highest d-excess and no stations existed in the northern part of Japan, i.e.,
ofprecipitationwasobservedintheTottoriPrefecturefacing the snowy regions of Tohoku and Hokkaido. The precipi-
theSeaofJapan.Theprecipitationsampleswerecollectedat tation isoscapes of the world (e.g., Bowen and Revenaugh,
theborderoftheTottoriandOkayamaprefectures(Hagaand 2003; van der Veer et al., 2009) exactly cover the Japanese
Katsuyama, unpublished data), and snowmelt recharges the archipelago and we can compare our data with their inter-
streamwaterinbothprefectures.Therefore,asYamamotoet polateddataproduct.However,thesestudiesbyBowenand
al.(1993)pointedout,thed-excessofthestreamwaterinthe Revenaugh (2003) and van der Veer et al. (2009) are based
OkayamaPrefecturewasrelativelyhigh(Table1). ontheGNIPdata.AsWassenaaretal.(2009)pointedout,the
Evaporationduringtheinfiltrationprocessesalsoaffectthe GNIPstationsareoftenspatiallydeficient.Therewereonly
d-excess value. The d-excess value will be depleted when 2 stations in Japan, and both stations were already closed.
theeffectsofevaporationarelarger.Evaporationfromafor- Therefore,theGNIPdataisinsufficientatthesmallcountry
est consists of canopy interception loss and evapotranspira- scale such as for Japan, even if the interpolation method is
tionfromtheforestfloor.Inourresult(Fig.5),thed-excess perfect.Ourresultsmayfillthespatialgapsinpreviousstud-
valueishigherattheSeaofJapanside.Kondoetal.(1992) ies.Inotherwords,althoughcomprisingonlyoneepisodeof
showedthatthecanopyinterceptionwaslargerattheSeaof sampling,ournationwidesystematicdatasetofstreamwater
Japan side compared to the Pacific Ocean side because the withhigherspatialdensity(Figs.1,4,5)revealsnotonlyspa-
canopyinterceptionwaspositivelycorrelatedwiththenum- tially integrated but also temporally integrated isotope sig-
berofprecipitationdaysinbothconiferousandbroadleaffor- nals of precipitation, and the stream water d-excess values
www.hydrol-earth-syst-sci.net/19/1577/2015/ Hydrol.EarthSyst.Sci.,19,1577–1588,2015
1586 M.Katsuyamaetal.:Spatialdistributionofoxygen-18anddeuteriuminstreamwaters
reflect the dampening of precipitation input, i.e., the differ- poral variations in isotopic signals when estimating, for ex-
encesinrainfall–runoffdynamicsamongthecatchments. ample, rainfall–runoff processes and/or the mean residence
time of stream water. Isoscapes of stream water reflect the
4.3 VulnerabilityofwaterresourcestoglobalWarming recharge processes from source water and the distribution
of water resources. Therefore, this technique will provide
As discussed above, winter snowfall recharges summer avaluablemethodforhydrologicalandecologicalresearch,
stream water at the Sea of Japan side, which experiences a andforpredictingtheimpactsofclimatechangeandestimat-
great deal of snowfall. This result means that the summer ingthevulnerabilityofwaterresourcesattheregionalscale.
water supply in this region is highly dependent on winter Inparticular,theseresultsfromJapan,acountrywithawide
snowfall.Brooksetal.(2012)clarifiedthatthewatersources rangeofclimaticandgeographicalconditionsacrossasmall
during summer depend on winter snow accumulated in the land area, represent a case study that will facilitate similar
mountains in western Oregon and pointed out the vulnera- studiesinotherregions.
bility of the system to the influences of a warming climate,
assnowpackvolumeispredictedtodeclineinthefuture.A
decreaseinsnowfallcausedbyglobalwarming,andthecon- TheSupplementrelatedtothisarticleisavailableonline
sequent vulnerability of water resources, are also predicted atdoi:10.5194/hess-19-1577-2015-supplement.
to occur in Japan (Kazama et al., 2008). Moreover, mean
air temperature is an important environmental parameter in
determining the δ18O and δ2H values of stream water, as
showninEqs.(6)and(8).Globalwarmingcanalsochange
Acknowledgements. Wewouldliketothankallofthecooperators
vegetation cover and evapotranspiration rates in the water-
oftheJWSM2003programfortheirintensivefieldsampling,labo-
shed, and the resulting amount of annual runoff (Gedney et
ratorywork,anddatabasecreation.Wealsogratefullyacknowledge
al., 2006). This implies that the importance of evapotran-
themanyresearcherswhokindlyofferedunpublishedprecipitation
spiration rates as a control parameter will change over the
data. This work was conducted as part of a Research Project at
longterm.Needlesstosay,transittimesdifferamongstream the Research Institute for Humanity and Nature (Environmental
waters.Therefore,theisotopicsignatureofstreamwaterre- ValuationProject),andpartlysupportedbyaJointUsage/Research
flects the history of precipitation in the catchment. In other GrantoftheCenterforEcologicalResearch,KyotoUniversityand
words,theisoscapesofstreamwaterrepresentanintegrated byJSPSKAKENHIgrantnumber25702020.
reflectionofthecontributionoftheseenvironmentalfactors,
and of the past and present climatic conditions. Therefore, Editedby:C.Stumpp
this kind of research should be conducted continuously ev-
ery few decades because the effects of climate change will
be reflected in these isoscapes over time and can provide References
information regarding changes in regional hydrological and
water-resourceconditions. Aggarwal, P. K., Alduchov, O. A., Froehlich, K. O., Araguas-
Araguas, L. J., Sturchio, N. C., and Kurita, N.: Stable isotopes
in global precipitation: A unified interpretation based on at-
mospheric moisture residence time, Geophys. Res. Lett., 39,
5 Conclusions
L11705,doi:10.1029/2012GL051937,2012.
Ahn,C.andTateishi,R.:Developmentofglobal30-minutegridpo-
Theδ18Oandδ2HisoscapesofstreamwaterinJapanshowed
tentialevapotramspirationdataset,J.Jpn.Soc.Photogr.Remote
clear spatial distributions which were effectively explained
Sens.,33,12–21,1994.
by three parameters: latitude, elevation, and mean annual Araguás-Araguás,L.,Froehlich,K.,andRozanski,K.:Stableiso-
temperature.Theseparametersarecommonlyknownascon- tope composition of precipitation over southeast Asia, J. Geo-
trol factors of the isotopes of precipitation. Therefore, our phys.Res.,103,28721–28742,1998.
datasetisapplicable,tosomeextent,asaproxyfortheiso- Bowen,G.J.andRevenaugh,J.:Interpolatingtheisotopiccompo-
topic composition of precipitation in Japan. This result re- sitionofmodernmeteoricprecipitation,WaterResour.Res.,39,
flects the advantage and importance of isotope contents in 1299,doi:10.1029/2003WR002086,2003.
Bowen,G.J.,West,J.B.,Vaughn,B.H.,Dawson,T.E.,Ehleringer,
streamwateratanationalscale,becausestreamwaterismore
J.R.,Fogel,M.I.,Hobson,K.,Hoogewerff,J.,Kendall,C.,Lai,
easilycollectablethanisprecipitation.However,thecompar-
C.-T. Miller, C. C., Noone, D., Schwarcz, H., and Still, C. J.:
ison of d-excess proved that the stream water d-excess val-
IsoscapestoAddressLarge-ScaleEarthScienceChallenges,Eos
ueswerebiasedtowardthevaluesofwinterprecipitationin
Trans.,90,109–116,2009.
snowy regions, although our sampling campaign was con-
Bowen,G.J.,Kennedy,C.D.,Liu,Z.,andStalker,J.:Waterbal-
ducted duringthe summer. Theseresults do notmerely sig- ancemodelformeanannualhydrogenandoxygenisotopedis-
nify the importance of continuously observing precipitation tributions in surface waters of the contiguous United States, J.
insnowyregions,butalsowarnustodiscreetlyusethetem- Geophys.Res.,116,G04011,doi:10.1029/2010JG001581,2011.
Hydrol.EarthSyst.Sci.,19,1577–1588,2015 www.hydrol-earth-syst-sci.net/19/1577/2015/