ARTICLE IN PRESSG Model ECOENG-2988; No.. ARTICLE IN PRESSG Model ECOENG-2988; No.. Location map of the study area, including the Heihe River Basin, the middle Heihe River Basin and the
Trang 1ARTICLE IN PRESS
G Model
ECOENG-2988; No of Pages 13
Ecological Engineering xxx (2014) xxx–xxx
ContentslistsavailableatScienceDirect
j o ur na l h o me pa g e :w w w e l s e v i e r c o m / l o c a t e / e c o l e n g
Sha Zhoua, Yuefei Huanga,∗, Bofu Yub, Guangqian Wanga
a State Key Laboratory of Hydroscience and Engineering, Tsinghua University, Beijing 100084, China
b School of Engineering, Griffith University, Nathan 4111, QLD, Australia
a r t i c l e i n f o
Article history:
Received 30 December 2013
Received in revised form 31 March 2014
Accepted 19 April 2014
Available online xxx
Keywords:
Middle Heihe River Basin
Land use change
Ecological environment
Arid areas
Environmental Kuznets curve (EKC)
a b s t r a c t
Withrapidsocio-economicdevelopmentoverthepastthreedecadesinChina,adverseeffectsofhuman activitiesonthenaturalecosystemareparticularlyseriousinaridregionswherelandscapeecologyis fragileduetolimitedwaterresourcesandconsiderableinterannualclimatevariability.Dataonlanduse, surfaceandgroundwater,climate,grossdomesticproduct(GDP)percapitafromthemiddleHeiheRiver Basinwereusedto(i)examinechangesinwaterconsumption,landusecomposition,andvegetation cover;(ii)evaluatetheeffectivenessofshort-termmanagementstrategiesforenvironmental protec-tionandimprovement,and(iii)applyandextendtheenvironmentalKuznetscurve(EKC)framework
todescribetherelationshipbetweeneconomicdevelopmentandenvironmentalqualityintermsofthe normalizeddifferencevegetationindex(NDVI).Theresultsshowedthatwithrapiddevelopmentof agri-cultureandeconomy,landusechangefortheperiod1986–2000wascharacterizedbytheexpansion
ofconstructedoases,considerablecontractionofoasis-deserttransitionalzoneandnaturaloases.This hasledtoadecreaseinecosystemstability.Since2001,effectivebasinmanagementhasbroughtabout improvedenvironmentconditions,withamoreoptimalhierarchicalstructureofvegetationcover.The originalEKCmodelcouldnotexplainmostoftheobservedvariationinNDVI(R2=0.37).Including addi-tionalclimatevariables,theextendedEKCmodeltoexplaintheobservedNDVIwasmuchimproved (R2=0.78),suggestingthatinclusionofbiophysicalfactorsisanecessaryadditionaldimensioninthe relationshipbetweeneconomicdevelopmentandenvironmentalqualityforaridregionswithgreat cli-matevariability.TherelationshipbetweenGDPpercapitaandNDVI,withtheeffectofprecipitationand temperaturetakenintoconsideration,wasadequatelydescribedbyanN-shapedcurve,suggestingthat therelationshipbetweensocietyandtheenvironmentfollowedaprocessofpromotion,contradiction, andcoordination
©2014TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-SA
license(http://creativecommons.org/licenses/by-nc-sa/3.0/)
1 Introduction
Thehumanactivitiesandclimatechangehaveinteracted
syner-gisticallytoimpacttherelationshipbetweensocialandecological
systemsinthelatetwentieth century(Steffenetal.,2005).The
complexinteractionsbetweensocialandecologicalsystemshave
fundamentallychangedinChinaduringthepastseveraldecades
Theimpactsofhumanactivitiesonnaturalecosystemare
espe-ciallyseriousinaridareaswherelandscapeecologyisveryfragile
duetolimitedwaterresources(LuoandZhang,2006).TheHeihe
∗ Corresponding author Tel.: +86 13911874076.
E-mail address: yuefeihuang@tsinghua.edu.cn (Y Huang).
RiverBasinisoneofthelargestaridinlandriverbasinsin north-westChina,whereoasesevolveasaresultofoppositeprocessesof oasificationanddesertification(Zhangetal.,2003;Suetal.,2007) Thelandscapecomposition,thespatialpatternordistributionof oases,desertandoasis-deserttransitionalzoneareknownasthe
‘eco-circlelevelstructure’,andthisnotionofastructureof eco-logical relevanceatalargescalecanbeusedasanindicatorof ecosystem stabilitytoidentifytheprocessesofoasification and desertificationbasedontherelativeabundanceofoases,desert andoasis-deserttransitionalzoneatthebasinorregionalscales (Zhang,2009,2010).Duringtheperiodfromthe1970stothe1990s, theecosystemchangedgreatlyintheHeiheRiverBasinbecauseof over-exploitationofwaterandlandresourcesforagriculturaland economicdevelopment,leadingtochangesintheeco-circlelevel http://dx.doi.org/10.1016/j.ecoleng.2014.04.020
0925-8574/© 2014 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-SA license ( http://creativecommons.org/licenses/by-nc-sa/3.0/ ).
Trang 2andseriouseco-environmentalproblems(WangandCheng,1999)
In2001,short-termmanagementprogramswereimplementedto
ceaseanynewlandreclamationforagriculturaluse,improve
agri-culturalwateruseefficiency,andconvertsomefarmlandbackto
forestandgrassland.Thisintegratedwaterresourcesmanagement
andregulationsystemhassofarbroughtaboutpositive
environ-mentaloutcomesfortheregion(WuandTang,2007).Therefore,it
isveryimportanttorecognizetheimpactsofhumanactivitieson
theeco-circlelevelstructureatabasinscaleandtheenvironmental
qualityatalocalscaleindifferentperiods,toidentifyandexplain
factorsandprocessesthatdrivetheenvironmentalchangeinthe
HeiheRiverBasin,andtoprovidesupportfordecisionmakingin
termsoflong-termstrategiesforenvironmentalprotection
Muchresearchhasbeenundertakentoinvestigatetheeco-circle
levelstructure,landuse/coverandenvironmentalchange,andthe
impacts of natural factorsand humanactivitiesin aridregions
basedonremotesensingandGIStechnology(e.g.Shoshany,2000;
Ayad,2005),especiallysincetheImplementationStrategyofthe
Land-UseandLand-CoverChange(LUCC)projectwaspublishedin
1995(Nunes andAuge,1999).InChina,researchhasbeen
con-centratedintheinlandaridareaofnorthwestChina(e.g.Luetal.,
2003;Zhaoetal.,2011).Usingamodelforinlandeco-circlelevel
structurewithwaterasthecriticalinput,Chenetal.(2004)
ana-lyzedmechanismforandcharacteristicsofchangesinecosystems
in aridregions, indicatingthat useof waterand landresources
wouldsignificantlyimpacttheecosystemstability.Thesignificant
positivecorrelationbetweenNDVI(normalizeddifference
vegeta-tionindex)andprecipitationdemonstratedthatclimatevariability
andchangecouldplayanimportantroleintheenvironment
vari-ability in aridand semi-aridregions (Liet al.,2003).However,
Kong etal (2010)investigated vegetationchangeand
environ-mentaldriversintheTarimRiverBasin,andtheresultsindicated
that environmentalfactorsonly contributed toa small
propor-tionofvegetation-relatedlandcoverchangeandtheinfluencesof
expandingagriculturalactivitieswerethemaincausesoflandcover
changeinaridregions.Daietal.(2010)alsoshowedthatthenatural
vegetationchangewasinfluencednotonlybyclimatechange,but
alsohumanactivitieswhichsignificantlychangedtheplanted
veg-etationbasedoncorrelationanalysisofNDVIanddrivingfactors
overnorthwestChina
Alltheresearchhascontributedtoecologicalsystemstudyin
aridregionsofnorthwestChina.However,therewerefew
compre-hensivestudiesintheHeiheRiverBasin,especiallyinthemiddle
HeiheRiverBasin,whereadvancedirrigationagricultureand
inten-sivehumanactivitieshaveresultedinanover-exploitationofwater
andlandresources,leadingtoseriouseco-environmentalproblems
intheMiddleandLowerBasin.Toimprovetheenvironmental
qual-ity,short-termmanagementprogramswerefirstimplementedin
themiddleHeiheRiverBasin.Thus,themiddleHeiheRiverisoneof
theidealregionsintheinlandaridareatodemonstratethe
relation-shipbetweensocietyandtheenvironment.Theimpactsofhuman
activitiesontheecosysteminthemiddleHeiheRiverBasinarestill
largelyunclearandthereisalackofquantitativeanalysisofthe
interactionsbetweensocioeconomicandbiophysicalprocessesat
thebasinscale.Accelerationofurbanizationprocesshasledto
seri-ousdeteriorationoftheecosystemintheHeiheRiverBasininthe
latetwentiethcentury(Mengetal.,2005)andtheeco-environment
hasbeen greatly improvedwith implementationof short-term
managementprogramssince2001(Dingetal.,2011).Therefore,
itisimportanttoquantifytheeffectivenessofmanagement
strate-giestoamelioratetheadverseeffectsofhumanactivitiesonthe
ecosysteminthemiddleHeiheRiverBasin
TheenvironmentalKuznetscurve(EKC)presentsahypothetical
relationshipbetweeneconomicdevelopmentandenvironmental
outcome (Grossman and Kreuger, 1991, 1995; Chowdhury and Moran,2012).Whiletheeconomicdevelopmentiscommonly mea-suredintermsoftheincomepercapita,amultipleofindicatorsof environmentaldegradationsuchasthelevelofairorwater pol-lutionhavebeenusedasameasureoftheenvironmentoutcome fromeconomicdevelopment(ShafikandBandhopadhyay,1992; GrossmanandKreuger,1995).Therelationshipbetweeneconomic growthandenvironmentalqualitycanbeverycomplicated,and hasbeenasourceofgreatcontroversy(Shafik,1994).Moreover, theEKCmodelmayrepresentanN-shaped,aninverseN-shaped,a U-shaped,aninverseU-shapedorevenalinearrelationship(Canas
etal.,2003),showingmultiplerelationshipsatdifferentstagesof economicdevelopmentandatdifferentspatialscales(Chowdhury andMoran,2012).Thereisbroadempiricalsupportforthe exist-enceofEKCforvariouspollutionindicatorsorvegetationcover
toexplainthedevelopment–environmentrelationship(Fosterand Rosenzweig,2003;Shen,2006;JalilandMahmud,2009).Lietal (2013)tested the relationship betweenpopulationgrowth and vegetationcover in21citiesin GuangdongProvince,China,the resultsshowthatthereisalong-terminvertedN-shaped relation-shipbetweenpopulationgrowthandvegetationcover,indicating thatpopulationincreasewithurbanizationmayhaveanegative
orpositiveimpactonthevegetationcoveratdifferentstagesof developmentbecauseoftheintensivehumanactivities.However, thecurrentEKCmodelhasnotconsiderednaturalfactorsin deter-miningvegetationchangeasameasureofenvironmentalquality
Asthehumanactivitieshavemultipleeffectsontheenvironmentin themiddleHeiheRiverBasinindifferentstagesofeconomic devel-opment,andclimatevariabilityandchangearestronglycorrelated withvegetationvariabilityinthearidregion(Zhaoetal.,2011),we proposeanextendedEKCmodeltoincludeanthropogenicand nat-uralfactorsforageneralexplanationofthedynamicrelationship betweenregionaldevelopmentandenvironmentalqualityforthe middleHeiheRiverBasin
Ourresearchhadthreeobjectives.Thefirstobjectivewasto ana-lyzethechangesintheecosystemundertheinfluenceofintense humanactivitiesbasedonchangesoflanduseandtheeco-circle levelstructurebetween1986and2000inthemiddleHeiheRiver Basin;thesecondobjectivewastoassesstheeffectsofshort-term managementstrategiesintermsofthespatiotemporalvariations
ofvegetationcoverintheZhangye–Linze–Gaotaibasinasan exam-ple;andthethirdobjectivewastodevelopanextendedEKCmodel
toexplorethedrivingmechanismoftheenvironmentalchangesfor possiblefutureprojections,takingintoconsiderationbothnatural andsocial-economicfactors
2 Materials and methods
2.1 Studyarea ThemiddleHeiheRiverBasin,betweentheYingluoGorgeand ZhengyiGorgestreamgaugingstations,islocatedinthecentralpart
oftheHexiCorridor,between98◦20–102◦12Eand37◦57-40◦03N (Fig 1 witha total areaof 2.61×104km2 The middle Heihe RiverBasin hasa number of administrativedistricts, including GanzhouDistrict,GaotaiCounty,LinzeCounty,ShandanCounty, MinleCounty,apartofSunanCountyofZhangyeCity,Jiayuguan City,and Suzhou Districtof JiuquanCity Thestudy areahasa temperate continentalarid climatewithadequatesunlight and infrequentoccurrenceofprecipitation.Themeanannual precip-itationisonly140mmandmorethanhalfofitoccursinsummer months(May–September).Themeanpotentialevapotranspiration
inthe regionis about1000–2000mmyr−1 (Wangetal., 2007) Theareahasanunbrokenirrigationagriculturalhistorysincethe
Trang 3ARTICLE IN PRESS
G Model
ECOENG-2988; No of Pages 13
Fig 1. Location map of the study area, including the Heihe River Basin, the middle Heihe River Basin and the Zhangye–Linze–Gaotai basin.
Handynasty,andisthemaincommoditygrainbaseintheHexi
Corridor.Zhangye Cityhasbeen widely knownas the“Golden
Zhangye”sinceancienttimes(Mengetal.,2003).However,thepast
30yearshavewitnessedthemostrapiddevelopmentof
agricul-tureandeconomyinthemiddleHeiheRiverBasin,accompanying
withsignificanteco-environmentalchanges.From1980to2010,
thepopulationincreasedfrom1.37millionto1.92million,andGDP
increasedenormouslyfrom0.49billionRMBperannumto51.87
billion,anincreaseofmorethan100times(EditorialBoardofGansu
Yearbook,1981,2011)
2.2 Datasourcesandprocessing
Inthisstudy,mostofthedatasetswereprovidedby
Environ-mentalandEcologicalScienceDataCenterforWestChina,National
NaturalScienceFoundationofChina(http://westdc.westgis.ac.cn),
includingtheGIMMSAVHRRNDVI(Normalizeddifference
vege-tationindex)(1982–2006),SPOTVEGETATIONNDVI(1998–2008),
landuse/coverdataoftheHeiheRiverBasinin1986and2000,daily
streamflowdataoftheYingluoandZhengyiGorgegaugingstations,
groundwaterdatain themiddle HeiheRiverBasin.In addition,
meteorologicaldata(precipitationandtemperature)were
down-loadedfromChina MeteorologicalData SharingService System
(http://cdc.cma.gov.cn),andeconomicdata(GDPandpopulation)
inGansuprovincewerederivedfrom“ComprehensiveStatistical
DataandMaterialson60YearsofNewChina”
TheGIMMSAVHRRNDVIproducts,at15-dayintervalwith8km
groundresolution,wereprocessedtoobtainanannualtimeseries
ofNDVI,whichwereusedtodeveloptheextendedEKCmodel.The
SPOTVEGETATIONNDVIproducts,at10-dayintervalwith1km
groundresolution,weretransformedintovegetationcover,using
themethodofvegetationfractionestimation(Li,2003).Then,the
maximum,average,andgrowthseasonaveragevegetationcover
wascalculatedtoanalyzetheeco-environmentalchangesinthe 2000s
Thelanduse/landcoverdatawereprocessedusingGISto ana-lyzelandusepatternsin1986and2000.Moreover,theeco-circle level structure wasanalyzedonthebasis of thelandusedata Naturaloasesincludeforestland,shrubbery,sparsewoodlot,high covergrassland,mediumcovergrassland,lake,permanentglaciers, beachesandflats,andwetland;constructedoasesincludeother woodland, irrigationcanalsand ditches,reservoirs,pond,urban land,ruralsettlementandotherlandforconstruction;oasis-desert transitional zonemainlyreferstolow cover grassland;andthe desert includes sandy land, Gobi, saline-alkali land, bare land, exposedrock,shingleland
TheannualstreamflowsoftheYingluoGorgeandZhengyiGorge gauging stationswereaccumulatedfromdailydata.Theannual precipitation data were derived from daily precipitation data andaveragedfromfourmeteorologicalstations(Zhangye,Gaotai, Jiuquan,Shandan),theannualtemperaturedatawerespatialand temporalaveragesprocessedfromthedownloadedtemperature data.TheannualGDPdatawereaccumulatedfromGDPdatafor individualadministrativedistrictsinthemiddleHeiheRiverBasin 2.3 TheEKCmodelanditsextension
TherelationshipbetweenGDPpercapitaandsomemeasure
ofenvironmentalqualityisknownastheenvironmentalKuznets curve(EKC).Aspreviously explained,this relationshipbetween economicgrowthandenvironmentalqualityisnotmonotonicand maypresentdifferentshapes.Acubicfunctioncanbeappliedto describethecomplicatedrelationship,andtheparametervalues associatedwiththecubicfunctiondefinetheshapeofthecurve (Martınez-Zarzosoand Bengochea-Morancho, 2004) Generally, GDPpercapitaistheindependentvariable,asameasureof eco-nomicgrowth,andthedependentvariablecanbediverse,including
Trang 4Table 1
Parameter values of the environmental Kuznets curve and the implied relationship
between environmental quality and economic growth.
indicatorsofenvironmentalpollutionorvegetationstatus,suchas
pollutionemissionsorvegetationcover.Moreover,thevariables
canbepretreatedwithlinearornaturallogarithmtransformations
(Sternetal.,1996),theyarealleffectiveinsupportingtheexistence
of theEKCbetweeneconomicdevelopmentand environmental
quality.TheoriginalEKCmodelisgivenby:
Eit=˛i+ˇ1
P
it+ˇ2
P
2
it+ˇ3
P
3
it+uit (1.a)
or
lnEit =˛i+ˇ1ln
P
it+ˇ2
ln
P
2 it
+ˇ3
ln
P
3
whereErepresentsthepollutionemissions(e.g.atmosphericCO2,
SO2,NOxemissionsandwastewaterdischarge)percapitaor
vege-tationcover;PstandsforthetotalpopulationandGDPisthegross
domesticproduct.Thespecificationisusuallyestimatedonpanel
datawithireferstothedifferentregionsandtreferstothe
differ-enttime,˛iistheindividualspecificinterceptofregioni,anduit
isastochasticerrorterm(AuciandBecchetti,2006).The
parame-tersˇ1,ˇ2andˇ3definetheshapeofthefunctionalrelationship
betweeneconomic growthandenvironmentalquality(Table1)
(Songetal.,2008),andtheshapecanvarydependingonselection
ofthestudyperiod,thestudyarea,andindicatorsofthe
environ-mentalquality.The3shapefactorsdependontheeffectofGDP/P
onenvironmentalequalityduringacertainperiod.Wemayexpect
thatthesignofˇ1,ˇ3tobepositiveandˇ2tobenegativewhen
theEKCexists.However,theeco-environmentinthestudyarea
wasaffectedbybothsocialandbiophysicalfactors,andthemodel
aboveonlyconsiderssocialfactors,i.e.theGDPpercapita,butdoes
notincludeanybiophysicalfactors.Wehypothesizedthatinthis
aridenvironment,biophysicalvariables,suchasthetemperature,
precipitation,streamflowcouldbejustasimportantindetermining
theenvironmentalqualityasmeasuredbyNDVIinthispaper.Thus,
weextendedtheEKCmodeltoincludeclimateandwaterresources
variables.Thesevariablesareassumedtoaffectthegrowthof
veg-etation,NDVI,hencetheenvironmentalquality.TheextendedEKC
modelforthemiddleHeiheRiverBasincanthusbewrittenas:
NDVIt =˛+ˇ1
P
t+ˇ2
P
2 t
+ˇ3
P
3
t +ˇ4CLMt+ˇ5WRt+ut (2)
whereCLMtisaproxyforclimatefactors,suchasannualaverage temperature;WRtisaproxyforwaterresources,suchasannual precipitationorstreamflow;andtheotherparametersarethesame
asdescribedabove.AsweconsideredthemiddleHeiheRiverBasin
asasingleregionforthepurposeofEKCmodeling,thesubscript,i,
isnolongerrequired
Furthermore,wemayhighlighttheeffectsofhumanactivities
onthevariationinNDVIbyremovingtheeffectsofnaturalfactors fromEq.(2).WecanthenderiveanadjustedNDVI(−)asfollow:
NDVI(t−) =NDVIt−ˇ4CLMt−ˇ5WRt=˛+ˇ1
P
t
+ˇ2
P
2
t +ˇ3
P
3
t+ut (3)
TheturningpointsofeconomicdevelopmentintermsofGDP percapitacanbecomputedas:
1= −ˇ2−
ˇ2−3ˇ1ˇ3
3ˇ3 , 2=−ˇ2+
ˇ2−3ˇ1ˇ3
3ˇ3
Iftheparameterssatisfyˇ1>0,ˇ2<0,ˇ3>0,theEKCreveals
anN-shapedrelationship(Table1 and1istheturningpointof localmaximumNDVI,2theturningpointoflocalminimumNDVI Thethreestagesseparatedby1and2describetheperiodswhen GDPpercapitachangeshavedifferenteffectsonhowNDVIvaries witheconomicdevelopmentwiththeadjustedNDVI(−).Asaresult, theeco-environmentalchangesandtheirdrivingmechanismfrom humanactivitiescouldbeaccentuatedandexaminedspecifically fordifferentperiodsofdevelopment
3 Results and discussion
3.1 Spatiotemporaldistributionofwaterresources CharacteristicsofandtrendsintheannualflowsoftheHeihe Riverwereanalyzed usingthestreamflowdatarecordedatthe YingluoGorgeandZhengyiGorgegaugingstationsovera54-year period(1957–2009).ThetwostationsseparatethemiddleHeihe Basinfromitsupperpartandlowerpart,respectively(Fig.1).Asthe precipitationoverthemiddlepartoftheHeiheRiverBasinisvery low,thereductionintheannualflowbetweenthetwolocationsis broadlyrelatedtotheannualwaterabstractionandconsumption alongthemiddlereachoftheHeiheRiver(Nianetal.,2013).In thispaper,thedifferenceintheannualrunoffvolumewastakento approximatetheannualwaterextractionfromsurfacerunofffor consumptioninthemiddleHeiheRiverBasin
The long-term average surface water consumption was 6.02×108m3 perannumoverthe54years,or37%ofthemean annualstreamflowattheYingluoGorge.AscanbeseeninFig.2, therewasasignificantincrease(p-value<0.01)ofsurfacewater consumptioninthemiddleHeiheRiverBasininthe1980s, averag-ing6.9×108m3perannum,orabouttwiceashighasinthe1970s Theincreasingtrendcontinuedintothe1990s,upto8.0×108m3
perannum.Inthe2000s,thesurfacewaterconsumptionstarted
toleveloffwithsomevariationsatabout6.9–9.3×108m3 annu-ally.Correlation betweenthestreamflowsattheYingluoGorge
Table 2
Changes in the eco-circle level structure from 1986 to 2000 (Ecological area includes total oases and oasis-desert transitional zone).
Natural oases (km 2 )
Constructed oases (km 2 )
Total oases (km 2 )
Oasis-desert transitional zone (km 2 )
Desert (km 2 )
Ecological area/total area
Oases/ecological area
Constructed oases/oases
1986 and 2000
Trang 5ARTICLE IN PRESS
G Model
ECOENG-2988; No of Pages 13
Fig 2.Time series of annual streamflow of the Heihe at Yingluo Gorge (upstream) and Zhengyi Gorge (downstream), and the difference to approximate the transmission loss and water abstraction in the middle Heihe River Basin (1957–2010).
andtheZhengyiGorgeshowsthatthestreamflowattheZhengyi
Gorgehasdecreasedsteadilyasa resultofextractionofsurface
runoffforconsumptivewateruse Forthesameannual
stream-flowattheYingluoGorge,theannualflowattheZhengyiGorge
decreasedbyabout2.5×108m3from1957–1979tothe1980s,and
byabout1.5×108m3fromthe1980stothe1990s(Fig.3)
How-ever,notonlyhasthestreamflowattheZhengyiGorgeincreased
inthe2000scomparedtothatinthe1990s,butalsothe
correla-tionbetweenannualstreamflowsatthetwogaugestationsbecame
stronger(R2=0.92forthe2000s,andR2=0.83forthe1990s)(Fig.3)
Strongercorrelationandtighterrelationshipbetweenthesetwo
streamflowstationsindicateamoreregulatedandmanagedsystem
forwaterabstractioninthe2000s.Thisoccurredlargelybecause
ofashort-termmanagementstrategyforwaterresourcesplanning andallocationforthemiddleHeiheRiverinthe2000s(Wangetal.,
2004)
In additionto the increasedwater abstraction from surface runoff, concurrentextractionofgroundwaterresourceshasalso beenincreasingsincethe1980s.Thedecliningtrendof ground-waterlevelsintheGanzhoudistrictisclear,wherethedepthto groundwaterat7monitoringstationsshowspersistentincrease ranging from5.32–12.85m(Fig 4).Groundwaterresources are depleting,andtherateofdepletionhasacceleratedinrecentyears Fortheperiodfrom1980to1992,therateofdepletionvariedfrom
Fig 3. Relationship between annual streamflows of the Heihe at Yingluo Gorge (upstream) and Zhengyi Gorge (downstream) (1957–2010).
Trang 6Fig 4. Annual depth to groundwater at 7 monitoring stations in the Ganzhou district, and the annual streamflow at the Yingluo Gorge (upstream).
0.09myr−1to0.31myr−1amongthese7stations;forthe13years
since(1992–2004),theratedepletionhasincreasedto0.31myr−1
to0.89myr−1
Inthe1990s,traditionalfloodirrigationwasthemainirrigation
method,waterwasdeliveredmostlythroughmaincanals,branch
canals,tertiarycanals,thewaterdeliveryefficiencywas0.65and
the water use efficiency was35% (Feng et al., 2000), meaning
thatmuchoftheirrigatedwaterwaslostfromthedelivery
sys-tem,becauseofpoorconstructionandmaintenance.Therewere
somany water storages whose utilizationefficiency were only
40–60%,andhalfofthewaterinwaterstorageswaslostdueto
leak-ageandhighrateofevaporation.Thus,accelerateddevelopmentof
agricultureandsocialeconomyhadledtoacontinuedincreaseof
waterconsumptioninthemiddleHeiheRiverBasin,and
concur-rentdecreaseintheavailablewaterresourcesforthelowerreaches
oftheHeiheRiver.Theshort-termmanagementprogramswere
implementedin2001whenmoderncanal-liningtechniqueswere
introduced,advanced water-saving irrigationmethods adopted,
suchassprinklerirrigationanddripirrigation.Theuseof
mod-erntechnology, theintegrated planningand allocationofwater
resourceshave significantlyincreasedwater useefficiency,and
ensuredappropriatewaterallocationbetweenthemiddleandthe
lowerreaches
3.2 Landusechangefrom1986to2000
TheintensivehumanactivitiesinthemiddleHeiheRiverBasin
areclearlyevidentnotonlyfromthewaterresourcesutilization,
butalsofromtheexploitationoflandresources.Thetotalareaofthe
studyareais2.16×104km2.Followingthearearatiofromlargeto
small,land-usetypesinthemiddleHeiheRiverBasinwereunused
land,grassland,arableland,forestland,aquatorium,and
residen-tialarea(residentialland,industryandmining)in1986,andthe
proportionsofthefirstthreewere53.0%,20.4%,18.7%.From1986
to2000,thearablelandareaincreasedby247.5km2,from18.7%
to19.9%;atthesametime,thegrasslanddeclinedby195.9km2,
from20.4%to19.5%.Analyzingthelanduseconversionbetween thesixlandusetypes(Fig.5 thetransferfromgrasslandtoarable landwasthelargest,about226.4km2,accountingfor90%ofthe lossofgrasslandand75%oftheincreaseinarableland.Inaddition, therewas40.6km2fromaquatoriumtoarableland,accountingfor 99%ofthelossofaquatorium(41.0km2).Apartfromtheexpansion
ofarableland,therewasasharpincreaseof20.9km2inthe res-identialareafrom387.2km2to408.1km2,andabout70%ofthis increasecamefromarableland
Thelanduseconversioncausedchangesintheeco-circlelevel structure,whichincludesthefollowinglanduses:naturaloases, constructedoases,oasis-deserttransitionalzoneanddesert Eco-logicalareainthisstudyincludesalltheaboveexceptthedesert, andtheratioofecologicalareatototalstudyareaindicatesthe frangibilityoftheecosystem,i.e.thesmallerthisratio,themore fragileandvulnerabletheeco-environmentis.Theoasesto ecolog-icalarearatioandtheconstructedoasestooasesratiorepresentthe relativestabilityofoasesandconstructedoases,respectively Dur-ingtheperiod1986–2000,constructedoasesincreasedby275km2
(6.0%),andtransitionalzone,naturaloasesdecreasedby176km2
(5.5%),90km2(3.8%),respectively(Table2).Theeco-environment showedadeclinetrendintherelativestabilityduetothesignificant expansionofconstructedoasesandanoticeabledeclineofthe tran-sitionalzone,withtheconstructedoasestooasesratioincreased from0.658to0.680andtheoasestoecologicalarearatiorisedfrom 0.685to0.703
Withtherapidagriculturalandeconomicdevelopment,arable landexpandedattheexpenseofsurroundinggrassland,andpart
of the arable land encroached for construction and residential development,resultinginanexpansionofconstructedoasesand considerablecontractionoftheoasis-deserttransitionalzone Dur-ingthesameperiod,forestlandandaquatoriumdecreasedbecause
ofmechanizedoperationsforfillingditchandreclaiminglandfrom lakemarshesonalargescale,thus,naturaloaseswereaffected.Asa resultofintensehumanactivities,theevolutionoflandusepatterns canbesummarizedastheexpansionofconstructedoases, consid-erablecontractionoftheoasis-deserttransitionalzoneandnatural
Trang 7ARTICLE IN PRESS
G Model
ECOENG-2988; No of Pages 13
Fig 5.Land use change from 1986 to 2000 in the middle Heihe River Basin.
Trang 8Fig 6. Time series of the vegetation cover from 1998 to 2008 in the Zhangye–Linze–Gaotai basin.
oases,andtherewasageneraldecreaseintheoveralleco-system
stability
3.3 Eco-environmentalchangesin2000s
Asshort-termmanagementprogramswerefirstimplemented
intheZhangye–Linze–Gaotaibasinintheearlier2000s,andthe
vegetationcovercanbeusedanindicatoroftheeco-environmental
qualityin thearidarea, weconsideredanumber ofchanges of
the vegetationcover tomeasure the improvement in the
eco-environmental quality in the Zhangye–Linze–Gaotai basin The
maximumvegetationcoverhasincreasedsteadilyby0.0063per
annum,andtheaveragevegetationcoverduringthegrowthseason
showsaparallelincreaseat0.0056perannum(R2=0.74)(Fig.6)
Thespatiotemperalchangesinthevegetationcovershowedthat
theenvironmentimprovedmostnotablyintheGanzhouDistrict (Fig.7 wherethechangesofvegetationcoverduringthegrowth seasonwere0.03–0.3inmostregions,andthevegetationcover increasedby5–60%.However,thevegetationcoverofsomearea
inthenorth(transitionalzones)andalongtheHeiheRiver(arable land)decreasedby0.1–0.3,or5–60%
Analyzingthevegetationcoverchangesduringthegrowth sea-sonfromMaytoOctober,theincreaseinthevegetationcoverwas notableforallmonthsduringthegrowthseasonexceptMay, espe-ciallyafter2001.Theincrease inthevegetationcover occurred
inAugustandSeptemberintheZhangye–Linze–Gaotaibasinwas
sodramaticthatsince2001,thepeakvegetationcoveralternated betweenJulyandAugust,while,thevegetationcoverinJulywas consistentlyhigherthanthatinAugustpriorto2001(Fig.8).The maximummonthlyvegetationcoverreacheditspeakinJuly2007
Fig 7. Spatiotemporal changes in the average vegetation cover from May to October (the growth season) in the Zhangye–Linze–Gaotai basin (1999–2007).
Trang 9ARTICLE IN PRESS
G Model
ECOENG-2988; No of Pages 13
Fig 8. Time series of the monthly vegetation cover from May to October (the growth season) in the Zhangye–Linze–Gaotai basin (1998–2007).
at0.39,andthiswasconsiderablyhigherthan0.33recordedfor
1998,and0.29for2001
Theannualmaximumvegetationcoverwasusedtoclassifythe
degreeofvegetationinthestudyareaaccordingtothefollowing
scheme:
• Verylow—0–15%
• Low—15–30%
• Mediumtolow—30–45%
• Medium—45–60%
• Mediumtohigh—60–75%
• High—75–100%
Ascan be seen clearly fromFig 9, there was a remarkable
increaseofthehighvegetationcover,increasingfromlessthan
1%in1998tonearly17%in2008.Thesuddenincreaseofthehigh
vegetationcoverfrom1999to2000wasattributedtothe
vegeta-tioncoverimprovementofforestlandandhighcovergrassland.At
thesametime,thevariationsinthelowandverylowvegetation
coverclassesseemedtooccurintandem(Fig.9).Changeinthelow
classhadacorresponding,almostequal,andoppositechangein
theverylowclass,suggestingthattheverylowclasswasregularly
transformedtolowclassandviceversaonalargescale.Allinall,
thehierarchicalstructureoftheannualmaximumvegetationcover
improved,andtherehavebeenimprovementsineco-environment
quality
3.4 TheextendedEKCmodel
Fig 10 shows the annual time series of (a) NDVI; (b) GDP
percapita;(c)averagetemperature;and(d)precipitationinthe
middleHeiheRiverBasin.ItisclearfromFig.10thattherehasbeen
ahugeincreaseinGDPpercapitafrom1982to2006,especially
inrecentyears.Againstthisconsistentandsustainedincreasein
economicdevelopment,wealsonotefromFig.10theconsiderable
interannual fluctuations in NDVI and climate variables These short-termvariationsfundamentallynecessitatedanextensionof theEKCmodeltoincludesomebiophysicalfactorstoexplainthe variationsintheeco-environmentalqualityasindicatedbyNDVI ParametersfortheoriginalEKCmodel,i.e.Eq.(1.a)wereestimated usingNDVI andGDPpercapitadatafortheperiod1982–2006 The relationshipbetweenNDVI andGDPper capitashowedan N-shaped relationship,andtheparameters satisfyˇ1>0, ˇ2<0,
ˇ3>0intheoriginalEKCmodel(Table3)
TheoriginalEKCmodelwithoutconsideringofnaturalfactors failedtofullyexplain mostof thevariationsinNDVIwithaR2
valueof0.37only(Fig.11(a)).TheextendedEKCmodel,i.e.Eq.(2), throughintroducingnaturalfactorscouldexplaintheinterannual variationsinNDVImuchbetterthantheoriginalEKCmodel.The resultoflinearcorrelationanalysisindicatedthatprecipitationand temperatureweresignificantlycorrelatedwiththeNDVI(r=0.56 forprecipitation;r=0.47fortemperature).Changesinprecipitation andtemperaturearecloselyrelatedtothewaterandenergy con-ditions,whichareimportantfactorsinvegetationgrowthinarid areas,especiallyforgrassandothervegetationtypesinthe oasis-deserttransitionalzone,whererunoffisrare,andsurfacewater
solimitedthatprecipitationdictatesthevegetationgrowthtoa greatextent.Thus,theextendedEKCmodelbelowwasselectedand parametersre-estimatedwithmultipleregressionanalysisthrough introducingnaturalfactorsasfollows
NDVIt =˛+ˇ1
P
t+ˇ2
P
2 t
+ˇ3
P
3
t+ˇ4TEMt+ˇ5PREt+ut (4)
where TEMt is theannual averagetemperature,and PREt isthe annual precipitationin themiddle HeiheRiver Basin.Allother parametersarethesameasthoseinEq.(2).Estimatedparameter valuesandmodelperformanceindicatorsarepresentedinTable3 forboththeoriginalandextendedEKCmodels
Trang 10Fig 9. Changes in the percentage area of the six vegetation cover classes in the Zhangye–Linze–Gaotai basin (1998–2008).
Fig 10.Time series of the NDVI, GDP per capita, average temperature and precipitation in the middle Heihe River Basin from 1982 to 2006.
Eq (4) describes therelationship between eco-environment
qualityanditsdrivingsocioeconomicandbiophysicalfactors
Com-paring theregressionresultsusing theoriginalEKCmodel, the
extendedEKCmodelisabletoexplainmostoftheobserved
vari-ations in NDVI for the study area (R2=0.78) The discrepancy
between theobserved NDVI and the modelledNDVI using the
extendedEKCmodelisnoticeablysmallerthanthatusingthe
orig-inal EKCmodel(Fig.12).Ifconsideringonly theclimatefactors
(precipitationandtemperature),thevalueofR2wouldbereduced
to0.57withthesamemultiplelinearregressiontechnique.Thus, bothhumanfactors,i.e.GDPpercapitaandnaturalclimate fac-torsarerelevantandneededtoexplaintheinterannualvariations
inNDVI inthemiddleHeiheRiverBasin.In addition,Eq.(4) is significantatthe1%level, soarethecoefficients forindividual termsinvolvingGDP/P,PRE,andTEMat the5%level Removing theeffectsofnaturalfactors,theN-shapedrelationshipbecomes strongerandtheamountofscatteraroundtheKuznetscurveis smaller(Fig.11(b)).TurningpointsintermsofNDVIintheEKC