人工衛星データ活用のための東アジアの植生調査

人工衛星データ活用のための東アジアの植生調査

著者 村本 健一郎

雑誌名 平成17(2005)年度科学研究費補助金 基盤研究

(B)(2) 研究成果報告書

巻 2002 2005

発行年 2006‑03‑01

URL http://doi.org/10.24517/00034745

第穏章 東軍ジア生態学国際会議（臣風陀S∋  

圏EAFESInternationalCongressinMokpo，Korea  

EastAsianFederationofEcologicalSocieties（EAFES），20−240ctober2004  

−EcologicalesearchandMonitoringinEnvironmentalChanges−  

Session：Friday（October22）  

M宜岳B岳亙a色岳⑳椚⑳柏em⑳竜e＄e閃＄日閃g督⑳『m⑳閑適色⑳面g⑳官vege亀a竜岳⑳閃6的a閃ge   Time：14：00〜17：00   

Room：AlO2  

0rgamizers   

−Ken−ichiroMuramot6，KanazawaUniversity（muramoto＠t．kanazawa−u．aC．jp）   

−Kyu−SungLee，InhaUniversity（ksung＠inha．ac．kr）   

−DafangZhuang，ChineseAcademyofSciences（zhuangdf＠1reis．ac．cn）   

一NaotoKamata，KanazawaUniversity（kamatan＠kenroku．kanazawa−u．aC．jp）  

瑠．Eeol⑳91Ca＝認esearcha閃d聞￠nit￠ri咽岳nEnv竜rommen飽帽ha閃9e＄盲neh巨na    DafangZhuang，JiyuanIJiu  

2。匿抒ec朗venes＄⑳督va雨MS Ve9e飴色摘n石門d竜6eS紬r帥e e＄亀邑ma軸門0官℡即㊤S竜Cam叩y  

Tae・KeunXiml，Kyu−SungLeel，andSeung−HoLee2  

乱 U＄eO督電imes即iesMOD】＄Lea官Arealndexd】a飴色Om⑳nitor竜◎mPeraせe紬帽＄竜   

KimS．H．1，S．H．Shinl，K．C．Jangl，K．S．I，eelandS．H．Lee2   乱 闘appin9L◎a官Å陀aEndexMSir19閥O四日＄da臨岳門6h岳na   

RonggaoIJiul，Jing．M．Chen2，JiyuanLiul，DafangZhuang■1   5．La閃de⑳VeF6月a＄＄i甘妃a竜岳⑳n and F⑳柑StÅ帽a Gha門9e De七ec竜ion随㌻四囲Z u＄En9   

Remo随一y＄e拍＄ed8mages   

J．H．Chun＊，J．H．LimandJ．H．Shin  

臥 日nc盲denceofJ叩aneSeOakWi陀inRe6a軸nを07bpo9F叩れyand＄un＄h盲ne   

R．KoMURAl，YIGETA2，A．M．LIEBHOLD3，K．EsAK［4，K．MuRAMOTO2，N．ⅨAMAm2  

1・Ec⑳00g8Ga＝竜℡＄earChand掴0閃岳電orin9inEnvironm㊧n短柑hange＄岳nehina  

DafangZhuang，JiyuanLiu  

lnstituteofGeographicalSciencesandNatura［ResourcesResearch，ChineseAcademyof  

Sciences，Beijing，China   Email：Zhuangdf＠lreis．ac．cn  

7brrestrialecosystemisakeycomponentofourplanet，butithasbeenchanglngbythe  

naturalandanthropoice鮎cts，SuChasdeserti且cation，urbanizationandglobalclimate  

change. It becomes more and more important to acquire the Terrestrial ecosystem 

StatuS andits changeinibrmation．Tb understand suchissues relating with the  

ecosystemanditschange，aSyStemtOmOnitortheecologicalsystemwas払undedin  

IGSNRR，CAS，WhichincludingCERN，ChineseResourcesandEnvironmentdatabase，  

remotesensingreceiverandprocessingsystemandecologicalsimulationmodellibrary．  

TheChineseEcosystemResearchNetwork（CERN），OneOfthefoundingmembersofthe   InternationalLongTbrmEcosystemResearchNetwork（ILRER），andGlobalTbrrestria1  

0bservation System（GTOS），WaS eStablishedin1988．CERN consists of36field  

researchstationsforvariousecosystems，includingagriculture，forestry，graSSlandand  

equlPPed variousinstruments to acqulre diffbrent ecosystem parameters，SuCh as  

realtimefluxdata，meteOrOlogicaldataandvegetationgrowthparameters，Thesedata，  

SOmearerealtime，WereCOllectedandqualitycontrolinitssynthesiscenter．Chinese   ResourcesandEnvironmentdatabasewasconstructedfrom1990，s，nOWhasincluded  

diHbrentresolutiongeospatialdatathatcontributedtoecologicalresearch．Thesedata  

includethemultitempora11andcoverdata，eCOlogicalsettingsdataand丘eldssample  

data・TheremotesenslngdatareceiverandprocessingsystemcanreceiveMODISdata  

andprocessthesedatatoextractecologicalparameters．Themodellibraryhascollected  

VariousmethodsreiatedwithecolOglCaisimuiation．The壬burcomponentsintegration  

WillimproveusinunderstandingtheChineseterrestrialecosystemanditschange．  

2．臣汗℡6七岳v◎n◎S＄⑳管var岳⑳u＄∀e9e食品竜岳⑳ni門diGe＄官⑳r肋e e＄せ毒ma鮎門⑳官紬FeS七Ca閃叩y   StruCture  

Tae−KeunKiml，Kyu−SungLeel，andSeung−HoLee2  

1■nhaUniversity，DepartmentofGeoinformaticEngineenng   2ForestryResearchLnstitute  

g2021540＠inhavision．inha．ac．kr  

Severalvegetationindices，Obtainedfrommulti−SPeCtralremotesensingdata，havebeen   widelyusedtoindirectlyestimatecanopystruCtureParameterSaCrOSSdi鮎rentbiomes．  

Althoughvegetationindiceshaveshowncertainrelationshipwithseveralbiophysical  

Variablesofvegetationcanopy，SuChasleafareaindex，biomass，andvegetationcoverage，  

theire鮎ctivenesshasnotbeenfu11ycon鮎medovertheverydensetemperatehrest  

Vegetation．Further，SOmeOfvegetationindiceshavelimitationtodescribe董brestcanopy   StruCtuTedue七otheinnuenceofbackgroundsoilandatmosphericattenuation．Inthis   Study，WeareattemPtingtoanalyzethee董艶ctivenessofvariousvegetationindicesover  

theforestecosystemhavingratherclosecanopycondition．  

Two vegetationindices，including normalized di鮎rentialvegetationindex（NDVI），  

simpleratio（SR），andenhancedvegetationindex（EVI），WereCOmParedinrelationtothe   leafareaindex（L射）0VerthestudyareaneartheSeoulmetropolitanarea．TheKyongan  

WaterShedcoveTSatOtalareaof561km20fconiferousplantationandnaturalstandsof  

mixeddeciduousspecies．DuringthegrowlngSeaSOnOf2003；fieldsurveywasconducted   OVer30groundsampleplotsto measure LAI，Stand density，and stand height were   measured．Each plot has an area of20Ⅹ20m2andincludes鋭ve subplotsforI．Al   measurement withinit．Plotlocations were determined uslng a di鮎rentialglobal   positioning sys七em（GPS）．Field measured LA王values were then compared with  

VegetationindicesthatwereextractedhlOmSatelliteremotesenslngdata，Whichwere  

Obtainedabou右上hesameseasonwiththeaeidsurvey．しapaDllltyOfvariousvegetation  

indicestoestimateLAIovertheclosecanopyforestecosystemarediscussed．  

3・Useo官t蓋meser岳esMOD日SLeafÅ陀abndexc8a飴色Om⑳n岳t即をempe柑短音即￠＄t  

KimS・H・1，S・H・Shinl，K・C・Jangl，K・S・LeelandS．H．Lee2  

1■nha University，■ncheon，Korea；2Forestry Researchlnstitute，Seoul，Korea．  

g2032128＠inhavision．inha．ac．kr  

AsonetheprimaryEOSsensors，MODIShasprovidednotonlyradiant−fluximagedata   but also severalbiophysicalvariables，includingleaf areaindex（LAI），fraction of   photosyntheticallyradiation（FPAR），andnetprimaryproduction（NPP），thatcanbe   ValuablesourceofinL）rmationforecologicalstudiesinrangeoftimeandspatialscales．  

Since2000，MODIS LAIdata（every eight−days composite）have been available王br  

Wholelandareasoftheearth．Theobjectivesofthisstudyaretoanalyzethepotential  

OftimeseriesMODISIJAIproducttomonitortheregionalscaleforestcoverchanges  

TimeseriesMODISI」AIdataoverthestudyreglOnWereObtainedfortheperiodof2000  

to2003・Asaninitialapproachtoassessthe qualityofthe data，We analyzedthose  

refbrence data that are delivered with LAIdata andinclude the LAIes七imation   algorithmusedandcloudcoverageforeverypixel．MeanLAIvalueovertheKoTean  

peninsulashowedveryunusualpatternfTOmJunetoAugustforeveryyear，Whichwas   CauSed by the highcloud coverage during the summer season．Those LAIvalues   estimated by the backup algorithm of using NDVIshowed ratherlow quality．  

There丘げe，itisstronglyadvisedtouseonlythoseLAIvaluesestimatedbythemain  

algorithmusingthecanopyradiativetransfbrmodelandnota鮎ctedbycloudcover．  

LAIofNorthKoreashowedlowerLAIthanSouthKoreaforspring，fallandwinter，but   Simi1armeanLAI（4．5Ⅰ．Al）duringgrowingseasonofsummer，Whichimpliestheshort   growingperiodduetotheclimatepatternbetweennorthandsouthKorea．Time・Series  

MODIS W data have shown the possibility of detecting and monitoring of fire  

damagedfbrests．ComparisonamongthreeyearsⅣLUUISl．AIproduct，WeWereabieto  

See highI」Alvalueforacertainyearwhenthe precipitationandtemperaturewere  

betterthantheotheryears．TimeseriesMODISLAIproductcanbeusefultomonitor  

SeVeralbioticandabioticchangeSOftempera七efbrestinregionalscale．  

乱闘appon9Lea官A帽alndexu引n9掴ODロ＄da臨斎nehi閃a  

RonggaoLiul，Jing・M・Chen2，JiyuanLiul，DafangZhuangl  

1lnstituteofGeographica［SciencesandNaturalResourcesResearch，ChineseAcademy  

2DeparEmentofGeography，UniversityofToronto，Toronto，Canada   Email：Ljurg＠Ireis．ac．cn  

Leafareaindex（IJAI）isoneofthesurfaceparametersthathasimportanceinclimate，  

Weather，andecologicalstudies．IthasbeenextensivelyinvestigatedtoestimatetheLAI  

fTOmremOteSenSlngmeaSurementS，butitisstillachallengetomapLAlinalarge  

reglOnalscale．TherearemainlytwomethodstoestimateLAI＆omsatellitedata，Oneis  

basedonthestatisticalrelationshipbetweenIJuandremotesensingbandswhichis  

OnlysuitabletolocalreglOnalscale，andanotheris鈷・Omtheinversionofphysicalcanopy   radiationtranSf白rmodelwhichismoreoptimisticforlargescaleLAIestimation．Inthis   presentation，aneWI」AIestimationmethodfFOmMODISdataisproposed．SinceLAIis  

akeyparameterindescriptionofvegetationstructure，thesatellitebandsdatamaybe  

di蝕ren七重brsameI」AIfr0血difhrentviewandsensorangles．Thecoupleofviewand  

SenSOranglewithsatellitebandsdatamaylmPrOVetheaccuracyofLAIestimation．The  

methodis based on the5−Scale modeldeveloped byJing．M．Chen．MODIS bands  

radiance were firstlysimulated to createlookup table uslngthe5−Scale modelwith  

inputofI」AI，Viewangle，SenSOrangleandlandcover−relatedparameters．ThentheI．AI  

WaSretrievedfromMODISbands datauslngSimulationlookuptable．Thealgorithm  

WaSaPPliedtoproduceI」AImapscoverlngallChinauslng16−dayscloud−freeMODIS  

data・Tovalidatethealgorithmandtheseproducts，ibursites，includingChangbaishan，  

Heihe，IJipingandXingguo，Were Selected董brvalidationundeTthe support ofCIDA   PrQjectsin2003．The丘eldmeasurementsusethecommercialTracingRadiationand   Architecture of CanopleSしrl 屁AU）．Landsat警hematlC朗apper trl M）scenes at3O−m  

resolutionatthesameperiodwereusedtolocategroundsitesandtofhcilitatespatial  

SCalingtol−kmpixels．ItisshownthattheaccuracyofLAIvaluesofMODISwasmore  

than80％・Randomandbiaserrorswerebothconsiderable．Biaswasmostlycausedby  

5．』andcover elass岳管岳6ati￠n岳nd F⑳re＄竜Å柑a 6h抑ge E》e始C鮎n菅⑳r P聞置u引関   配em⑳モe8y＄en＄eddmage＄  

J．H．Chun＊，J．H．LimandJ．H．Shin  

Korea Forest Researchlnstitute，＃207，Chongnyangni2−dong，Dongdaemunqgu，Seou1   130−712，RepublicofKOREA  

Chunjh69＠orgio．net  

Thisstudyaimstocharacterizepresentlandcoverandtodetectitschangeovertimefor  

theDemi1itarizedZone（DMZ）anditsvicinityofKoreausingremotelysensedimages．  

DMZ，a4kmwide（N−S），248kmlong（W−E）corridorshapedareaalongthestalemate  

partlywe11conservedhasbeenwitnessingperiodicaldisturbancecontinuouslydueto  

the armyoperations and artincial丘re clearingthe sight．The area surrounding the  

DMZhasbeenunderthepressureforurbanizationandtheexpansionofcropland．IJand  

COVerChangesinDMZanditssurroundingareaweremappedby？OmParingpastand  

presentI」andsatTM5and7satelliteimages，Andlandcovermapsshowing7categories  

Ofpresentland cover andits change were developed using supervised classification  

techniquesincludingcalculationofNDVIandGISoverlaybetween1987−1989and2001．  

Themaincategoryofquiteabigchangeis−Forestlclass−ofwhichthestemdensityand  

theconversionofforestsintocroplandandbarrenland董brfoodproductionandsupplyof  

fuelwood．Andwhatisworse，mOStOfthecroplandsdevelopedonthemountainslopes  

in North Korea are known to be experiencing increased erosion and decreased soil 

f6rtilityApartfromthequality，鈷restsinsidetheDMZshowedbetterconnectivitythan  

theecosystemofDMZanditsvicinityhaveoccurredanddeterminingwheretoconserve  

andrestore丘）rmaintainingthestructureandfunctionofDMZecosystem．  

6・8ncid◎門CeO官Japa閃e＄◎⑳akWi陀i門Re日at旨⑳閃電07軸⑳9ra田内y品田d＄u門＄h岳門℡  

R・KoMURAl，YIGE1712，ArM・LJEB＝OLD3，K．EsAK［4，K．MuRAMOTO2，N．KAMAm2；  

1−shikawaNationalCor■egeoflbchnoJogy，鴨ubata，lshjkawa，Japan・  

2KanazawaUniversity，Kanazawa，lshikawa，Japan・  

3NortheasternResearchStation，USDAForestService，Morgantown，WVUSA，  

4lshikawaForesiExperimentStation，lburugi，Ishikawa，Japan・  

Emaii：komura＠ishikawa−nCt．aC．jp  

Japaneseoakwi1t（JOW），CauSedbyanambrosiafungus勉励eha quezdT70）7aand   VeC七oredbyasymbioticambrosiabeetlel％ち坪uSqueZd陀rUS（Murayama），hasbeen  

PreValentinJapan董brthelas七色fteenyears．InIshikawaPrefbctuTe，incidenceofJOW   WaS鮎stfoundinthewesternportionoftheprefectuTein1997・In七hisstudy，We  

investigated the relationship between historicalJOWincidence with topographical  

fbatures，including slope，aSPeCCt and exposure to solarradiation・The studyarea  

COnSistedof1520Ⅱ1Ⅹ1380mreglOninIshikawaandFukniprefbctures．Thesitewasa   mountainousregionincludingMt．KariyasuandwasnearlyuniformlycoveredbyR）reSt   Stands dominated by6？ueⅣuS CEゴ如uh．．Killed trees wereidentified by their red  

COlorationinaerialphotographstakenbeforenaturalfallcolorchangeinmiddleof  

October・Theslopeandaspectwerecalculatedfromadigitalelevationmodel（DEM）．  

Solaranglefrom6ÅMto8AMonlstAugustwerecalculatedandrelativeamountsof  

SOlar radiation at each positionwasalso calculated usingthe DEM．Alldata were  

COmpiled using a geographicinformation system（GIS）．Theincidence ofJOW was   higherineast−fhcingslopesthanthatinwest−fhcingslopes．Pquezdt7WuSadultsare  

POSitively phototactic．The east−払cing slopes receive more solar radiationin the   morning，Whenadult且ightsoccur．IncidenceofJOWtended七obegreaterinlocations  

thatTeCeivedmoresolarradiationinthemorming・  

7．Jap忍門e＄e⑳a随W岳陀＄pF℡adana8yzedwi肋6−＄and掴u陀i。＄Pa竜岳a日・＄Ca8eDa忠a  

R．KoMURAl，A．M．LIEBHOLD2，K．EsAK13，K．MuRAMOTO4，N．KAMAm4；  

1lshikawaNationalCollegeoflt，Chno．ogy，lbubata，［shikawa，Japan・  

2NortheasternResearchStation，USDAForestService，Morgantown，WV，USA・  

3−shikawaForestExperimentStation，T5urugi，■shikawa，Japan・  

4KanazawaUniversity，Kanazawa，■shikawa，Japan・  

Emaii：komura＠ishikawa−nCt．aC．jp  

Japaneseoakwilt（JOW）hasbeenknowntoexistinJapansincethe1930s．Themost  

earlyoutbreaskofthis diseaserecordedwereonlyafbwyearsindura七ionandwere  

confinedtoafbwareasontheJapanSea（westeTn）coastofJapan，butinthelastten   yearsepidemicshaveintensi且edandspreadtotheisland swesterncoastalareas．The  

Symbiotic ambrosiafungus励彪elea queldvon2is the causalagent ofoak dieback  

associatedwiththedisease，anditisvectoredbyl％訪pusque］壷l7nruS（Murayama）．  

Thisisthe鮎stexampleofanambrosiabeetlefungusthatkillsvigoroustreeS．Previous  

StudieshavedocumentedthehistoricalspreadofconcurrenciesofJOWattheTeglOnal  

SCalelevelanditwasrecognizedthatJOWspreadbuttrendsofthisexpansionhavenot  

beenanalyzedstatistically．Inthisstudy，Statisticalanalysisofthespreadoftheoak  

dieback was conducted using GIS tooIs and expansion behavior was quantified  

atdiffbrentspatialscalew．There appeartobeatleastthreediffbrentscales atwhich  

expansionoffocioperate．Theseresultssuggestthatadultbeetleshavecababilitiesof  

movingovervariousspatialscales．  

凱＄tudy⑳閃La関東−u＄◎化and−GOV即eha門gea門dl鞄r㌻e軸ia日6a抽0閃Gye−e旨閃eh岳閥  

JiyuanLiu，QuanqinShao  

lnstituteofGeographicalSciencesandNaturalResourcesResearch，ChineseAcademyof   Sciences  

Email：Liujy＠igrnss．ac．cn  

TheinterestintheterrestrialcaTboncycleatnationallevelshasbeenincreasingsince  

the agreementofKyotoProtocoIwasreached．Carboncyclelinkshuman，biological，  

geochemicaland atmospheric process．The change of the magnitude and spatial   distributionofnetcarbonnuxesbetweenecosystemsandtheatmosphereiscontrolled  

mainlybyprlOrdisturbanceandland−uSe／1andcoverchange．Tbaddressthescientific  

issuessuchastemporalandspatialpatternOfterrestrialecosystemcarbonsink，and  

driving mechanism and scenarios ofcarbon cycle，etC．，a method ofgeo−information   SCiencefor studying carbon cycle ofterrestrialecosystemsis proposed．Bottom−up   approach and top−down approach are combined by means of scaling models．The  

bottom−uP aPPrOaChis based on observations of comprehensive network ofcarbon  

StOrage and carbon cycle process ofterrestrialecosystems，adaptive experiments of  

biologicalpTOCeSSeS，andresearchesoncarbontransportationprocessesofrivers．The  

top−downapproachisbasedondetectinglandcoverchange andretrievlngeCOlogical  

ParameterS byusing satellite．Retrievalmodels ofcarbonbudgets are developedby  

informationofgeologicalprocessesandecologicalprocesses．Onthebasisofanalyzing  

data・at−pOints collected by observation stations of Chinese Ecosystem Research   Network，Stations of Chinese Forest Ecosystem Research Network，and stations of   ChinaFLUX，COmbined with the retrievalmodels，a numericalsimulation modelof   terrestrialecosystemcarboncycleis constructedbymeans ofsurface theorem，grid  

generationmethodandgridcomputingtechmique．Patternandprocessofcarboncycle  

are to be simulated；naturaland humanimpacts on carbon cycle of terrestrial  

ecosystemsaretobeanalyzed；andevolutiontrendsofcarboncycleprocessofterrestrial  

ecosystemsalletObediscussedundertheconditionofglobalclimatechange．