characterizing and evaluating the arctic digital elevation model … · 2017. 6. 7. ·...

Post on 25-Jan-2021

1 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

  • Arjan J.H. Meddens1*, Lee A. Vierling1, Jan U.H. Eitel1, Natalie T. Boelman2, Jyoti Jennewein1 & Andrew Maguire1

    2a. Methods

    Datasets:•ArcticDEM (5mresolutionproduct)

    •AlaskanairborneLiDARdataset(black,Fig.2,Hubbardetal.2011)

    Studyareas:Selectedsamplelocations(red,Fig.2)ofthe2011LiDARdataset:

    §Location1:Arctictundravegetation,intermediatetopographicalcomplexity

    §Location2:BlackandwhiteSpruceforest(open- closed),lowtopographicalcomplexity

    CharacterizingandevaluatingtheArcticDigitalElevationModelproductwithLiDARdataforspatialmodeling

    1. Introduction

    • TheArcticDEM isapublic-privateinitiativetoproducedigitalsurfacemodelacrosstheArctic(Fig1).

    • TheDEMisautomaticallycreatedbyusingopticalstereohigh-resolutionimageryat2mand5mspatialresolution(Noh&Howat,2015)

    • CurrentlyArcticDEMsforAlaskaandpartsofCanada,Norway,Russia,andIcelandhavebeenproducedandareavailablefordownload

    Objective:CharacterizetheaccuracyandnatureoftheArcticDEMdatasetforselectedregionsinAlaskawhereairborneLiDARisavailable

    Acknowledgements: WewouldliketothankMikeWulder andJoanneWhiteforhelpfulcomments.ThisresearchwassupportedbyNASAABoVE projectsEitel-01andBoelman-01andtheUniversityofIdahoCollegeofNaturalResources.

    References: Hubbard,T.D.,Koehler,R.D.,andCombellick,R.A.(2011).High-resolutionlidar dataforAlaskainfrastructurecorridors,inDGGSStaff,ElevationDatasetsofAlaska:AlaskaDivisionofGeological&GeophysicalSurveys,291p.Noh,M.J.,&Howat,I.M.(2015).Automatedstereo-photogrammetricDEMgenerationathighlatitudes:SurfaceExtractionwithTIN-basedSearch-spaceMinimization(SETSM)validationanddemonstrationoverglaciatedregions.GIScience &RemoteSensing,52,198-217

    2b. Methods

    AnalysisI: CompareelevationheightsoftheArcticDEM totheLiDAR-derivedbareearthandvegetationsurfacemodelsacrosstheselectedsamplelocations(using400randomlyselectedpixelsataminimumdistanceof100mfromeachothertoaccountforspatialautocorrelation)

    AnalysisII: AssesseffectsofvegetationheightintheArcticDEM databycomparingatransectoftheArcticDEM,LiDARbareearth,andLiDARvegetationheightmetrics

    Fig 2. Extent of the 2011 LiDAR data set (black) and the two selected sample locations (red).

    Fig1. Illustration of the proposed ArcticDEMdomain. Note that only parts of the domain have currently been processed (source: pgc.umn.edu/arcticdem).

    Fig. 3. Comparison between the ArcticDEM (x-axis) and LiDAR (y-axis) products (Location 1: top, Location 2: bottom; comparison with the LiDAR-derived bare earth: left and LiDAR-derived vegetation surface model: right). Data are 400 randomly selected pixels > 100 m apart.

    3a. Preliminary results (Analysis I)• ThecomparisonsbetweentheArcticDEM andtheLiDARdatasetsshowverygoodagreement(R2s>0.99,RMSEs<7 m,andbiases<5m

    • Thecomparisonsofthevegetationsurfacemodels(rightpanels)aremoreaccuratethanthebareearthmodelswithreducedRMSEsandbiases

    • MorecomplextopographydoesnotreducetheoverallaccuracyoftherelationshipsbetweentheArcticDEMandtheLiDAR-derivedheightmetrics(notshown) 4. Conclusions

    • TheArcticDEM isauniquedatasetusefulfortopographicandecologicalmodelingamongmanyotherapplicationsacrossthearctic-borealdomain(Fig.1)

    • TheArcticDEM datashowgoodagreementtoanindependenthighresolutionLiDARdataset

    • Usersshouldbeawarethatwhendense(forest)canopiesoccurinthedataset,theArcticDEM followsthetopofthecanopyratherthantheunderlyingtopographyofthelandscape

    Nextsteps:• Includemorelocationswithmorecomplextopographyandvegetationstructure• DevelopavegetationstructureproductfromtheArcticDEM foranimalmovementandtreeline modeling

    3b. Preliminary results (Analysis II)

    • Intallanddensevegetation,theArcticDEM followsthetopofthecanopyratherthantheunderlyingtopographyandgapsof~10marenotcaptured(Fig4.)

    Fig 4. Transect of 1400m from dense forest towards a more open vegetation valley bottom (see black line in the inset image). In black the ArcticDEM data is shown, in red the LiDAR-derived bare earth model, and in green the LiDAR vegetation height (with corresponding second y-axis), all units shown are in meters.

    1University of Idaho, Department of Natural Resources and Society, Moscow, ID (*ameddens@uidaho.edu) 2Lamont Doherty Earth Observatory, Columbia University, Palisades, NY.

    AK

    Location 1

    Location 2 Location 1

    Location 2 Location 2

    Location 1

    BareearthcomparisonVegetationheightcomparison

top related