ModellingtheLinkageBetweenLandscapeMetricsand

ProcediaEnvironmentalSciences13(2012)2131–21451878-0296©2011PublishedbyElsevierB.V.Selectionand/orpeer-reviewunderresponsibilityofSchoolofEnvironment,BeijingNormalUniversity.doi:10.1016/j.proenv.2012.01.202Availableonlineat(2011)2157–2171:AnAllometricModelM.Y.Wua,L.Xuea,W.B.Jina*,Q.X.Xionga,T.C.Aia,B.L.Lia,baEngineeringResearchCenterofWetlandAgricultureintheMiddleRegionofYangtzeRiver,(MinistryofEducationofChina),YangtzeUniversity,Jingzhou,Hubei434025,ChinabEcologicalComplexityandModellingLaboratory,DepartmentofBotanyandPlantScience,UniversityofCalifornia,Riverside,CA92521-0124,USAAbstractStudyingquantitativerelationshipsbetweenlandscapepatternandwaterqualityisafundamentalsteptoassesstheimpactsofnon-pointsourcepollution.Manyhydrologicalmodelswithmulti-functionalityhavebeendevelopedasusefultoolstostudyseveralkeymechanismsinnon-pointsourcepollution.Inlandscapeecologicalstudies,however,theempiricalmodellingapproacheshavebeendominatedwithemphasisontherelationshipsbetweenthelandscapemetricsandwaterqualityindices.Themaintechniquesfordevelopingthosemodelsoflandscape-waterqualityarestatisticalregressionanalysisbasedonlinearmodels.Inthisarticle,AllometricmodelsandthetraditionalmultiplelinearregressionmodelsforestimatingthelinkagebetweenlandscapemetricsandwaterqualityweretestedinSihuBasin,HubeiProvince,China.Themodelsatpatchclasslevelwereestablishedin24hydrologicalunitsofthebasin,whichtookninewaterqualityindices(EC,pH,SS,DO,COD,TN,TP,NO3--N,NH4+-N)asthedependentvariablesandeighteenlandscapemetricscalculatedinFRAGSTATS3.3asindependentvariables.Theresultssuggestedthat,comparedwiththetraditionalmultiplelinearregressionmodels,AllometricmodelsweremoresuitableforSS,DO,TP,TN,NH4+-N,inwhichlandscapepatternmetricscouldexplainthe80.5%,77.7%,58.2%,43.9%,67.6%oftotalvariation,respectively.TherehadlittledifferencebetweenmultiplelinearregressionmodelsandAllometricmodelsforECandNO3--N.ThecoefficientsofdeterminationinAllometricmodelswerenotasstrongasthatobtainedinthemultiplelinearregressionmodelsforpHandCOD.TheaboveresultsindicatedthatusingAllometricmodelmaypotentiallyprovideanewwaytostudythelinkagebetweenlandscapemetricsandwaterqualityindices,whichwillhelpprotectourregionalwaterresources.©2011PublishedbyElsevierLtd.*Correspondingauthor.Tel:[phone];fax:[phone].E-mailaddress:[e-mail].©2011PublishedbyElsevierB.V.Selectionand/orpeer-reviewunderresponsibilityofSchoolofEnvironment,BeijingNormalUniversity.2132M.Y.Wuetal./ProcediaEnvironmentalSciences13(2012)2131–21452158M.Y.Wuetal./ProcediaEnvironmentalSciences8(2011)2157–2171Keywords:wetlandagriculturebasin;landscapemetrics;waterqualityindices;AllometricmodelNomenclaturePLANDPercentageoflandscapeNPNumbersofpatchesPDPatchdensityTETotaledgeEDEdgedensityLSILandscapeshapeindexLPILargestpatchindexAREA_MNMeanpatchareaAREA_SDStandarddeviationinpatchareaAREA_CVCoefficientofvariationinpatchareaPARA_MNMeanperimeter-arearatioSHAPE_MNMeanshapeindexSHAPE_AMArea-weightedmeanshapeindexFRAC_MNMeanpatchfractaldimensionFRAC_AMArea-weightedmeanpatchfractaldimensionAIAggregationindexIJIInterspersion&juxtapositionindexCOHESIONCohesionindex1.IntroductionNon-pointsource(NPS)pollutionresultedfromtheagriculturalproductionthreatenedthewaterqualityandaquaticecosystems[1-3].Inthefactorscausedthenon-pointsource,landuseandlandcoverarepredominatedandnumerousstudieshavebeenconductedtherelationshipsbetweenlandusesandwaterqualitywithinwatersheds[4-7].However,thereisagrowingdemandforlarge-scalelandtransformation,whichaffectsstreaminavarietyofwaysacrossnumerousspatialandtemporalscales[8-12].Recently,scientistsinhydrology,ecology,geography,pedology,environmentalsciencesareconcernedaboutthechangesinlandscapecompositionofwatershed,theircumulativeimpactiononwaterquality,andemergeshundredsofwaterqualitymodelsonnon-pointsourcepollutionmechanismand2133M.Y.Wuetal./ProcediaEnvironmentalSciences13(2012)2131–2145M.Y.Wuetal./ProcediaEnvironmentalSciences8(2011)2157–21712159nutrientmigrationandtransformation.Ofparticularconcernisthedegreetowhichlandscapeconditionsatwatershedscalesinfluencenitrogen,phosphorus,andsedimentloadingstosurfacewaters[13,14].Highlevelsofnutrientsandsedimentinwatercanposesignificanthumanhealthandecologicalrisks[13].Inwatershedscalesmodels,therearethemechanismmodelsbasedonhydrologicalprocessesandempiricalmodelsbasedonthecorrelativeregressionanalysisbetweenlandscapeandwaterquality[15-21].However,itoftenoccursthatparametersofthesemodelshaveundefinedecologicalsignificancewhenwemodelslandscapeandwaterqualityusingthesemethods,whichistheonereasonthatleadstothelimitedapplyingofempiricalmodels.Atpresent,Allometricmodels,whichalsonamedmultiplepowerexponentregressionmodelsandoriginatedfromtherelationshipbe