Wetland-Watershed Modelling and Assessment: GIS Methods for Establishing Multiscale Indicators (pp. 231-250)
$100.00
Authors: (Javier Martinez-Lopez, M. Francisca Carreno, Jose Antonio Palazon-Ferrando, Julia Martinez-Fernandez, Miguel Angel Esteve, Departamento de Ecologia e Hidrologia, Universidad de Murcia, Campus de Espinardo, Murcia, Spain)
Abstract: In thecontextofwetlandecosystemmanagement,acombinationofapproaches
involvingdifferenttimeandspatialscalesmustbeapplied.Oneprimarydriverof
wetland degradationisagriculturalexpansionatwatershedscale.Wetlandshaveun-
dergoneseveralhydrologicalandbiologicalchangesasaconsequenceofincreased
water inputsfromagriculturaldrainageoffthewatershed.Fortheestablishmentof
suitable wetlandecologicalindicators,watershedscalestudiesfocusingonpressures
influencing ecosystemdynamicsarenecessary.Specificenhancedmethodsforwater-
shed modelling,wetlandmappingandlandcoverassessmentarethusessentialtools
for wetlandmonitoringandmanagement.
WatersheddrainingtotheMarinadelCarmolísemiaridwetlandinMurciaRegion
(SE Spain)wasdelimitedusingadigitalelevationmodel.Mapalgebraoperations
were appliedontheelevationmodeloftheCampodeCartagenacoastalplaintorein-
force existingdrainagenetworkandtoforceflowaccumulationfromalldrainingareas
around wetlandperimetertoconvergeintoasinglepointwithinthewetlandarea.Wa-
tershed delineationwasthusimproved.
A landuse/landcovermapoftheCampodeCartagenawasthenobtainedforyear
2008-09 bymeansofsupervisedclassificationofLandsatimages.Asetoffourspec-
tral indiceswerecalculatedandincludedintheclassificationanalysisusingacombi-
nation ofbandsinordertobetterdiscriminatevegetation,waterbodies,infrastructures
and baresoil.Anenhancedclassificationprocedurebasedonmaximumlikelihoodand
random reselectionoftrainareaswasapplied.Object-basedanalysisoftheLandsat
scenes basedonautomaticimagesegmentationdiminishedtheoccurrenceofisolated
pixelsintheclassification.Theproposedclassificationmethodologyshowedgreat
accuracy,thusimprovingtheresultsoftraditionalclassificationtechniques.
WetlandplantcommunitiesinMarinadelCarmolíwetlandweremappedin2008
by meansofremotesensingtechniquesusingsatelliteandairborneimages.Charac-
teristic plantcommunitieswerefirstcharacterizedbycombiningfieldworksampling
of plantspeciesandmultivariateanalysis.Georeferencedsamplingunitswerefur-
ther usedastrainingareasforsupervisedimageclassificationofplantcommunities.
Maps obtainedshowedgreataccuracy.However,sensorsareadequatefordifferent
applications.
The proposedsetofGISmethodologicaltoolscontributestoimprovethestudy
of wetlandplantasindicators,themappingandfuturemonitoringofwatershedland
coverclasses,andthestudyofwetlandplantcommunitychangesovertime.Allpieces
of softwareusedinthestudyarefreeandmainlyopensourceprograms,whichmake
it aninexpensiveanduniversalmethodology.