SatelliteRemoteSensingapplicationsforLandslideHazardMonitoring

Ashutosh Limaye,EricAnderson,BrianReevesNASA/SERVIRScienceCoordinationOffice

https://ntrs.nasa.gov/search.jsp?R=20170011111 2020-07-24T11:29:08+00:00Z

Outline

• BackgroundonSERVIR• NASA’sworkingloballandslidehazardmonitoring• Countrycasebriefsonlandslidemonitoringusingremotesensing• CasestudyonrelationshipbetweenfiresandlandslidesinNepal• CEOSLandslideDisasterWorkingGroupPilot• Areasforcollaboration

SERVIRisajointdevelopmentinitiativeofNASAandUSAID,workinginpartnershipwithleadingregionalorganizationsaround theglobe,tohelpdevelopingcountriesuseinformationprovidedbyEarthobservingsatellitesandgeospatialtechnologiestoaddressFood

Security,WaterandDisasters,WeatherandClimate,andLandUse/LandCoverChange.

Preventing seafood-borne illnesses in

Central America by mapping harmful

microalgae

Supporting food security in Nepal by

monitoring agricultural drought

Conserving forests in eastern and southern

Africa by mapping land cover and land use

change

Protecting lives in South/Southeast Asia by

monitoring and forecasting intense

thunderstorms

Helping herders and farmers in West Africa by

detecting ephemeral water bodies

TheCurrentSERVIRHubNetwork

SERVIR’sapproachtodisasterriskreduction

WaterandWater-relatedDisastersThematicServiceAreaofSERVIR

ü Shiftingfromproductcreationtoservicedesignanddelivery

ü Improvingscientificandtechnicalrigorofservicesthroughexternal“TechnicalAssessmentGroups”

ü BringingmoreinnovativeandappropriatesciencefromtheUS• 118USinstitutionsacrossallthematicserviceareas

ü Enhancingcollaboration acrossSERVIRhubs

SERVIR’sapproachtodisasterriskreduction

RoleofSERVIRinNASAEarthScienceDisastersProgram

üMatchneedsonthegroundwithtechnicalexpertisethatEarthscientistscanprovide

üBuildcapacityofagenciesaroundtheworldtouseEarthobservationinformation

üProvidefeedbacktoNASAontheutilityofscienceproductsfordisastermanagement

üProvideinputfromtheinternational“applications”communityperspective

NASA’sworkingloballandslidehazardmonitoring

• TheglobalLandslideHazardAssessmentforSituationalAwareness(LHASA) model isdevelopedtoprovidesituationalawarenessoflandslidehazardsforawiderangeofusers.[1]

• Considersweightedsatellite-derivedprecipitation(GPMIMERG),roads,deforestationandburning,tectonicfaults,bedrockconditions,andslope

• “GlobalLandslideNowcast”isupdateddaily

• NASAGloballandslidecatalog[2]• developedwiththegoalofidentifyingrainfall-triggeredlandslideeventsaroundtheworld,regardlessofsize,impactsorlocation

[1]Stanley,T.,andD.B.Kirschbaum (2017),Aheuristicapproachtogloballandslidesusceptibilitymapping,Nat.Hazards,1–20, doi:10.1007/s11069-017-2757-y.http://link.springer.com/article/10.1007%2Fs11069-017-2757-y

[2]Kirschbaum,D.B.,Adler,R.,Hong,Y.,Hill,S.,&Lerner-Lam,A.(2010).Agloballandslidecatalogforhazardapplications:method,results,andlimitations.NaturalHazards,52(3),561–575.doi:10.1007/s11069-009-9401-4.https://data.nasa.gov/Earth-Science/Global-Landslide-Catalog-Export/dd9e-wu2v

https://pmm.nasa.gov/applications/global-landslide-model

Casebriefs

1. 2009LandslidesinElSalvador,andfollow-onhazardsanalysis2. 2015Gorkha earthquakeinNepal3. AppliedresearchforbetterunderstandingoflandslidehazardsinRwanda

Casebrief:2009ElSalvadorlandslides

• ConvergenceofatropicalstorminthePacificandalowpressuresystemintheAtlanticledtoextremelyintenseandprolongedrainfall,andresultingfloodsandlandslides

Datafrom

disconnecteddecision

supporttoolsare

difficulttoassimilate

andcanprovide

conflicting

information

Masswastingsusceptibility

Laharinundationzone

(Anderson,2013)• Charteractivationinvolvingrapidresponsemapping• Value-addedproductssupportedreconstructionplan• Realizedthatfollow-onappliedresearchwasneeded

Casebrief:2015Gorkha earthquakeinNepal

• 4312landslidesidentifiedfrom10satellites:fewerlandslidesthanexpectedforanearthquakeofthismagnitude,possiblyduetomuchlessshakingatthesurface(Kargel etal.2016)

• Networkanalysisshowingvolunteerscientists&analysts(red)andconnectionswithuser/decisionmakingagencies(green)(Schumannetal.2016)

NASA/GSFC

Casebrief:AppliedresearchforbetterunderstandingoflandslidehazardsinRwanda

• FromUSGeologicalSurvey,weneedtoknow4thingsaboutlandslides1.Whenwilltheyhappen? 2.Wherewilltheystart?3.Wherewillthego? 4.Whatcouldbeaffected?

254landslideeventsidentifiedthroughvisualinterpretationofhighresolutionimagesinGoogleEarth

Preliminaryhazardmapderivedthroughlogisticregressiontesting(Piller andAnderson2015)

Possiblenextsteps:Considernewwaystocollectcrowd-sourceddata(e.g.,SpaceAppsChallenge)

Casestudy:Nepalfire-landsliderelationship

Overarchingquestion:CanwedetectanyrelationshipsbetweenfiresandlandslidesinNepal,asseenfromthesatelliteremotesensingperspective?

• Justification• Post-firelandslideprobabilityisoftenconsideredinU.S.GreatBasin• BurnedAreaEmergencyResponse(BAER)teamsassesspostfirethreatstolives,property,andresources

• Apparentlackofresearchintofire/landslidelinkagesinNepal

• FiresinNepal• ProlongeddryseasonsandlowerwinterprecipitationinNepalhaveincreasedwildfireincidences

• FireisamajorcauseofforestdegradationinNepal NASA Earth Observatory

(2016)

Casestudy:Nepalfire-landsliderelationship

• Studyareas:• Koshi Basin,Nepal– rainfall-triggeredlandslides(ICIMODdatabase)• Gorkha earthquakeaffectedareainNepal(Kargel etal2016)

• Researchquestions• Istherearelationshipbetweenfirefrequency/severityandlandslideoccurrence?

• Howdoestherelationshipchangewhenconsideringrainfall- vs.earthquake-triggered landslides?

Casestudy:Nepalfire-landsliderelationship

• Approach• Testresponseofrainfall-triggeredlandslides(ICIMOD/Koshi basin)andearthquake-triggeredlandslides(Kargel etal.2016)toseveralenvironmentalfactors,includingnormalizedburnratio(NBR)derivedbyLandsat7from2003to2015,usinglogisticregressionapproach

• Potentialexplanatoryvariables:

Variable Abbrev. Data Source Spatial Res Time SummaryNormalized Burn Ratio NBR LANDSAT 7 30 m 2003-2015 (SWIR-NIR)/(SWIR+NIR)Fire Occurrence Fires MODIS MCD45A1 500 m 2003-2015 (Fires)/(catchment)Drainage Density DD ALOS 5m DEM 5 m (str length)/(As)Topographic Wetness Index * TWI ALOS 5m DEM 5 m ln(As/tanβ)

Sediment Transport Index * STI ALOS 5m DEM 5 m (As/22.3)m (sinβ/0.0896)n

Stream Power Index * SPI ALOS 5m DEM 5 m AstanβPopulation Density Pop Dens Landscan 1 km 2010 (People)/(catchment)Height Above Nearest Drainage HAND ALOS 5m DEM 5 m Vertical distanceSlope Slope ALOS 5m DEM 5 m (rise)/(run)Euclidean Distance to Streams Eucl Str ALOS 5m DEM 5 m Straight line distanceAspect Aspect ALOS 5m DEM 5 m Direction of slopeProfile Curvature Prfl Crv ALOS 5m DEM 5 m Parallel to dir. max slopePlan Curvature Plan Crv ALOS 5m DEM 5 m Perpindicular to max slopeFlow Accumulation Flow Acc ALOS 5m DEM 5 m Accum. pixel x pixel flowCHIRPS CHIRPS CHIRPS Monthly 0.05° 2003-2015 Average monthly accum.CHIRP CHIRP CHIRP Monthly 0.05° 2003-2015 Average monthly accum.

As = surface area of catchment; β = slope in degrees; m = 0.6; n = 1.3(Moore et al, 1988)

Casestudy:Nepalfire-landsliderelationship

Results forearthquake-induced landslides:5%improvementconsideringNBR

NotconsideringNBR

ConsideringNBR

Predicted Accuracy0 1

Observed 0 304 223 57.70%1 125 594 82.60%

Overall Accuracy: 72.1%

Overall Accuracy: 77.3%

Predicted Accuracy0 1

Observed 0 363 211 63.20%1 118 601 83.40%

Y= −#$. &&' +(&. &*' ∗ ,-./0) +(&. &'' ∗ 23456)

Y= −##. 789 +(&. &*: ∗ ,-./0) +(&. &'7 ∗ 23456) +(#9. $78 ∗ ;<5)

Casestudy:Nepalfire-landsliderelationship

Results forrainfall-induced landslides:negligibledifferenceconsideringNBR

NotconsideringNBR

ConsideringNBR

Overall Accuracy: 59.4%

Predicted Accuracy0 1

Observed 0 564 426 57.00%1 378 613 61.90%

Overall Accuracy: 60.6%

Predicted Accuracy0 1

Observed 0 591 399 59.70%1 381 610 61.60%

> = −#. :&$ +&. &88 ∗ ,-./0 +(&. &&# ∗ 23456)

Y=−&. *9? +8. *:# ∗ ;<5 +&. &8' ∗ ,-./0 +&. &&9 ∗ 23456

Casestudy:Nepalfire-landsliderelationship

Knownlimitations

• Rainfall-inducedcase• Exactlandslidedatesunknown;therefore,cannottesttimingofexplanatoryvariables

• Earthquake-inducedcase• Onlyconsideredonemajortriggeringevent(2015Gorkha earthquake)

• Givenunknownspecificdatesofmostlandslidesinstudyset,wehadtoconsiderburningoveralongperiodoftime(versussingleburnevents)

• Furtherdatacollectionontimingandlocationofburning,triggerfactors(e.g.,rainfall,earthquakes),andlandslideeventscouldshedmorelightonfire-landsliderelationships

Formore,seeReeves(2017):https://ntrs.nasa.gov/search.jsp?R=20170001625

CEOSLandslidePilot

MaingoalsTodemonstratetheeffectiveexploitationofEarthobservations(EO)dataandtechnologiestodetect,mapandmonitorlandslidesandlandslidepronehillsides,indifferentphysiographicandclimaticregions.

ToapplysatelliteEOacrossthecycleoflandslidedisasterriskmanagement,includingpreparedness,situationalawareness,responseandrecoverywithadistinctmulti-hazardfocusoncascadingimpactsandrisks.

• Co-leads• Dr.DaliaKirschbaum,NASAGoddardSpaceFlightCenter,Maryland,USA• Dr.JonathanGodt,LandslideHazardsCoordinator,U.S.GeologicalSurvey,Colorado,USA• Dr.Jean-PhilippeMalet,SchoolandObservatoryofEarthSciences,UniversityofStrasbourg,France• Dr.SigridRoessner,GFZGermanResearchCentreforGeosciences,Germany

CEOSLandslidePilot

Threeobjectives(2016-2019):1. EstablisheffectivepracticesformergingdifferentEarthObservationdata(e.g.optical

andradar)tobettermonitorandmaplandslideactivityovertimeandspace.2. Demonstratehowlandslideproducts,models,andservicescansupportdisasterrisk

managementformulti-hazardandcascadinglandslideevents.3. Engageandpartnerwithdatabrokersandenduserstounderstandrequirementsand

userexpectationsandgetfeedbackthroughtheactivitiesdescribedinobjectives1-2.

Twomainfocusregions:NepalandthePacificNorthwestUnitedStates,includingWashingtonandOregon.

Plansfortheexperimentalregionsarestillindevelopment,butinclude:SoutheastAlaska,China,theCaribbean(HaitiandLesserAntilles),Peru,andIndonesia.http://ceos.org/ourwork/workinggroups/disasters/landslide-pilot/

Areasforcollaboration• CEOSLandslidesPilot– Earthobservationsfocus

• SERVIR:ADPCandHubConsortiummembersareconductingadditionalconsultationsandneedsassessmentswithstakeholdersintheregiontodesignfutureservices.Aretherewaystocollaboratewithotherinternationaltechnicalinstitutionstocollectivelyaddresslandslideriskmanagement?

• WeareinterestedinfeedbackandfindingwaystoconnectresearchandapplicationstobroaderNASAresources,includingfuture:

• NISAR– NASA-ISROSyntheticApertureRadarmission• SWOT– SurfaceWaterOceanTopographymission• Landsat9

• AGUFallMeeting2017sessionsNH018: Landslidedynamics:hazardandriskassessment,triggering,modeling,in-situobservations,andremotesensing:https://agu.confex.com/agu/fm17/preliminaryview.cgi/Session23681

Thankyou