SpacePolicyxxx(2013)1e13 ContentslistsavailableatScienceDirect Space Policy journal homepage: www.elsevier.com/locate/spacepol Earth observation from space e The issue of environmental sustainability Sylvie Durrieua,*, Ross F. Nelsonb aUMRTETISIrstea-Cirad-AgroParisTech/ENGREF,MaisondelaTélédétectionenLanguedoc-Roussillon,500,rueJ.F.BretonBP5095,34196MontpellierCedex 05,France bCode618,Hydrospheric&BiosphericSciencesLaboratory,NASA/GSFC,Greenbelt,MD20771,USA a r t i c l e i n f o a b s t r a c t Articlehistory: Remotesensing scientists work under assumptions that should notbe taken for granted and should, Received7February2013 therefore,bechallenged.Theseassumptionsincludethefollowing: Receivedinrevisedform 1.Space,especiallyLowEarthOrbit(LEO),willalwaysbeavailabletogovernmentalandcommercial 1July2013 spaceentitiesthatlaunchEarthremotesensingmissions. Accepted16July2013 2.Spacelaunchesarebenignwithrespecttoenvironmentalimpacts. Availableonlinexxx 3.MinimizationofType1error,whichprovidesincreasedconfidenceintheexperimentaloutcome,is thebestwaytoassessthesignificanceofenvironmentalchange. Keywords: 4. Large-area remote sensing investigations, i.e. national, continental, global studies, are best done Spaceremotesensing Earthobservationsystems fromspace. Spacejunk 5.Nationalspacemissionsshouldtrumpinternational,cooperativespacemissionstoensurenational Sustainability controlanddistributionofthedataproducts. Sustainablemanagement Atbest,allofthesepointsarearguable,andinsomecases,they’rewrong.Developmentofobservational Lidar spacesystemsthatarecompatiblewithsustainabilityprinciplesshouldbeaprimaryconcernwhenEarth Forest remotesensingspacesystemsareenvisioned,designed,andlaunched.Thediscussionisbasedonthehy- Biomass pothesisthatreducingtheenvironmentalimpactsofthedataacquisitionstep,whichisattheverybeginningof Lifecycleassessment theinformationstreamleadingtodecisionandaction,willenhancecoherenceintheinformationstreamand strengthenthecapacityofmeasurementprocessestomeettheirstatedfunctionalgoal,i.e.sustainableman- agementofEarthresources.Wesuggestthatunconventionalpointsofviewshouldbeadoptedandwhen appropriate,remedialmeasuresconsideredthatcouldhelptoreducetheenvironmentalfootprintofspace remotesensingandofEarthobservationandmonitoringsystemsingeneral.Thisarticlediscussesthesefive assumptionsinthecontextofsustainablemanagementofEarth’sresources.Takingeachassumptioninturn,we findthefollowing: (1)SpacedebrismaylimitaccesstoLowEarthOrbitoverthenextdecades. (2) Relatively speaking, given that they’re rare event, space launches may be benign, but study is meritedonupperstratosphericandexosphericlayersgiventhechemicalactivityassociatedwithrocket combustionby-products. (3)MinimizationofTypeIIerrorshouldbeconsideredinsituationswhereminimizationofTypeIerror greatlyhampersorprecludesourabilitytocorrecttheenvironmentalconditionbeingstudied. (4)Incertainsituations,airbornecollectsmaybelessexpensiveandmoreenvironmentallybenign, andcomparativestudiesshouldbedonetodeterminewhichpathiswisest. (5)Internationalcooperationanddatasharingwillreduceinstrumentandlaunchcostsandmission redundancy.Givenfiscalconcernsofmostofthemajorspaceagenciesee.g.NASA,ESA,CNESeitseems prudenttocombineresources. (cid:1)2013PublishedbyElsevierLtd. 1. Introduction Until the middle of the 20th century, environmental sciences * Correspondingauthor.Tel.:þ33(0)467548732;fax:þ33(0)467548700. couldstillbebasedonleisurelymethodsofdatacollectionwhich E-mail addresses: [email protected] (S. Durrieu), ross.f.nelson@ nasa.gov(R.F.Nelson). werecompatiblewiththerelativelyslowspeedofenvironmental 0265-9646/$eseefrontmatter(cid:1)2013PublishedbyElsevierLtd. http://dx.doi.org/10.1016/j.spacepol.2013.07.003 Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 2 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 changes and with the scales of studies [1]. Presently, constantly resources, those measurement and monitoring systems should, improved and updated information is a given for resource moni- themselves,beassustainableaspossible.Inshort,thismeansthat toringandmanagementastheycancapturethedynamicnatureof theenvironmental,social,andeconomiccostofamissionmustbe environmentalconditionssuchasclimatechange,waterallocation, lessthan the corresponding returns. Referring toecological engi- aswellassoilandbiodiversityloss[2]. neeringprinciples[9,10],environmentalimpactsfromthesystem Asearlyasin1969,inthefirsteditorialoftheRemoteSensingof productionstagetoitsendoflifeshouldbebetterunderstoodand Environment journal, Simonett [1] stated that “the quickening of takenintoaccount.Thiswouldenablethedesignofmeasurement science, and resource use, and the demands of society have systems capable of providing valuable information for managers increased the urgency to obtain quantitative, timely information whileminimizingtheirinevitableenvironmentalimpacts.Forget- abouttheenvironmentatavarietyofscalesinspaceandtime”.He tingtolimitthoseimpactswhendesigninganobservationsystem posited that observations made using ground-based sensors, isliabletoleadtosuboptimaloreveninappropriatesolutions. aircraft,andspaceplatformscouldhelptomeettheseinformation Section 2 examines two issues that challenge the assumption requirements [1]. Three years later, the launch of the Earth’s Re- thatspaceborneEarthobservationsystemsaresustainable.Section sources Technology Satellite ERTS-1, later renamed Landsat-1, 2.1 calls into question the basic assumption that Earth-orbiting marked the beginning of the Landsat era, thereby providing sig- platforms will always be available to the civilian remote sensing nificant impetus for the development of environmental applica- community. Section 2.2 focuses on the environmental impacts of tions based on remote sensing data at local to global scales [3]. spaceactivityontheEarthandreportsonhowtheseimpactsaffect Some of the most common applications in the remote sensing sustainability. Section 3 presents some unconventional points of world, such as agriculture or water management, can be traced viewthat,inouropinion,arerequiredtoaddressthesustainability backtoresearchperformedonspecificlandscapefeaturesidenti- issueofspace-basedEarthobservationsystems.Inthissectionwe fiedbyKondratyevetal. [4]ononeof the first Landsat1 images also suggest using environmental life cycle assessment as an recorded in July 1972. Navalgund et al. [5] classified the current analysis tool that might be particularly relevant to help reaching remote sensing applications into the following categories: sus- thesegoals.Section4detailspossibleinitiativesthatfollownatu- tainableagriculture,watersecurity,environmentalassessmentand rallyifthesenon-traditionalpointsofviewaredeemedvalidand monitoring,disastermonitoringandmitigation,andinfrastructure that might be taken to mitigate impacts associated with space development. Other fields of research such as fisheries manage- missionsandimprovesustainabilityofEarthobservationsystems. ment, weather and climate studies have also benefited from the Weaddressedsomeissuesthatarecommontoallremotesensing development of the remote sensing sector [5]. More recently, missions.Inadditionastudycasewaschosen,i.e.vegetationlidar remotesensingdatahavemorebeeninstrumentalinadvancingthe missions,toillustrateafewcase-specificpossibleactions.Section5 fieldsofecology,biodiversityandconservation[6]. summarizesandconcludes. As environmental impacts of human activity make resource managementmoreandmorecomplexandas,atthesametime,our 2. WhatcanputthesustainabilityofEarthobservationfrom understandingofcomplexnaturalprocessesincreases,ourneedfor spaceinjeopardy? critical information layers at appropriate spatial and temporal scalesandextentsincreasestoo[2].Thegrowingnumberoftheme- In what follows, sustainability is considered with respect to specificsatellites,notedbyNavalgundetal.[5],reflectsatechno- durability, space debris, and with respect to space activity as an logical response that can help to overcome such limitations, Earthpollutionsource. therebyfacilitatingnaturalresourcemanagement. Therelevanceofsuchspacebornetheme-specificmissionscan 2.1. UncertaintiesaboutthedurabilityofEarthobservationfrom betakenasagivenfromthemeasurementpointofview.However, space the premise that spaceborne observation can best provide infor- mationforsustainablemanagementofEarthresourcesshouldbe 2.1.1. Historicalcontext,currentstateandoutlookforspaceactivity subjectedtomorecriticaldebate.IndeedthesustainabilityofEarth The development of the space sector began in 1957 with the observationfromspaceisnotasevidentasitmightseem,apoint launch of Sputnik, the first artificial, Earth-orbiting satellite. The thatisseldomdiscussed. total number of launches since 1957 exceeded 5000 during year Thepaper,whichispartiallybasedonapreviouswork[7]aims 2009(seeFig.1)andthemeanannualnumberoflaunchesoverthe totakeafreshlookatmeasurementprocessesdesignedtosupport tenlastyearshasbeenslightlyhigherthan65[11].Eventhoughthe themonitoringofEarthresourcesandtopromotedebateaboutthe numberoflaunchesperyearhavetrendeddownwardsincetheend roleofremotesensingfromspacewithinthecontextofsustainable of the cold war in 1991 (Fig.1), the total number of operational management of these resources. Sustainability is defined as the satellites has continually increased due to a rise in both the life- capacitytoendure,andsustainabledevelopmentas“development timeandmeannumberofsatellitesperlaunch. that meets the needs of the present without compromising the Thedevelopmentofspaceactivitieshaslongbeendrivenbythe abilityoffuturegenerationstomeettheirownneeds”[8].Inthis political and military aspirations of the USA and Russia, the two paper, we look at Earth observation sustainability from two mainplayersinthissector.Oneofthepeacedividendsfromtheend different directions. First, given the number of space launches to ofthecoldwarwastheriseincommerciallyviableapplicationse date, the amount of space debris currently in orbit, and the ex- e.g. telecommunication and Earth observation e and the emer- pectednumberoffuturelaunches,canwesafelyassumeaccessto gence of new space powers, which led to the whole-scale trans- orbitforoperationalenvironmentalmissionsinthefuture?Second, formation of the space sector. This transformation affected space givenafullaccountingoftheenvironmentalcostsassociatedwith programsbutalsospaceactivityarchitectureasawhole,affecting spacelaunches,aresatellitesnecessarilythebestwaytosustainthe bothmilitaryandcivilianapplications[12],pavingthewayforthe flow of measurements needed to monitor the status of Earth’s emergenceofnewfeatureswhicharespecifictothecurrentsetof environment? activesatellites.Therearecurrentlycloseto1000activesatellitesin The common thread of this paper is the idea that to increase orbit, operated by 41 countries and several international consor- efficiencyanddurabilityofobservationandmeasurementsystems tiums [13]. Fig. 2a and b shows the distribution of satellites ac- designed to support sustainable management of the Earth cording to orbit classes and scientific/commercial disciplines, Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 3 Annual and cumulated number of launches from 1957 to 2010 First artificial satellite launched end 1957 e) v es per yearm) 111102460000 456000000000 aunches (cur ber of launch(histogra 24680000 123000000000 d number of l m e Nu 0 0 ulat m 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 Cu Year Fig.1. Evolutionofthenumberoflaunchessince1957.Thehistogramrepresentsthenumberofannuallaunches(Yleftaxis)whilethecurverepresentsthecumulatednumberof launchessincethefirstartificialsatellitearrivedonorbit(Yrightaxis).ThisfigureresultsfromtheanalysisofdatafromMcDowell’sdatabase[11]. respectively.Ofthe135activeEarthobservationsatellites,120are trendtowardsthedevelopmentofbothmicro-satellitetechnology onLEO.Theprofoundchangesinthespacesectorledtoareduction etherebymakingspacetechnologymoreaffordablefordeveloping inpublicinvestmentsthathaveweakenedthespacesector[12],at countrieseandthedeploymentofmulti-satelliteconstellations.It leastwhenusinglaunchactivityasayardstick.However,according canbeassumedthatallthesefactorswillresultinanongoingin- toPasco[12],projectsthatbringspacetosocietyratherthanthe creaseinthenumberofactivesatellites. reverse,suchliketheEuropeaninitiativefortheGlobalMonitoring forEnvironmentandSecurity(GMES),recentlyrenamedCoperni- 2.1.2. Spacedebrisandthreatstofutureorbitalactivities cus, could bolster this sector. Furthermore, there is a noticeable Spacedevelopmenthasresultedinanincreaseintheamountof space debris to such an extent that orbital debris is currently a threattospacecrafthealthandsafety[14].Spacedebrisismadeup a) ofnon-functionalsatellites(23%),upperstagesoflaunchers(18%), functionaldebris(14%),e.g.bolts,belts,andfragments(45%)orig- Elliptical Orbit inating from collisions, launcher upper stages and spacecraft ex- 4% plosions. The current number of catalogued objects, i.e. objects larger than 5e10 cm at Low Earth Orbit (LEO) altitudes and 30e 100cmatGeosynchronousOrbit(GEO)altitudes,whicharetracked Geosynchronous Orbit Low Earth Orbit by the US Space Surveillance Network, is about 16 000 and is 48% 41% increasingbyseveralhundredsperyear[14,15].Estimationsofnon- cataloguedobjectsvarydependingonthesource.Accordingtothe French Space Agency (CNES) estimates, there are roughly w200 000 objects with sizes ranging between 1 and 10 cm and Medium Earth w35millionofbetween0.1and1cm[16]. Orbit Mostobjectsmakinguporbitaldebrispopulations,andaround 7% 40%ofdebrisgreaterthan1mminsize,islocatedinLEO[16](see b) Fig.3).LEOspacedebrismitigationisacriticalissueforspaceac- tivitysustainability.Uptonowfouraccidentalcollisioneventshave Earth alreadybeenrecorded[16]alongwiththreeothersuspectedcol- Others observation Technology lisions [17]. A collision between a satellite and a piece of debris 11% 14% Development largerthan10cmwouldleadtolossorexplosionofthesatellite.To 5% Navigation 8% prevent collisions involving catalogued debris, alert systems for Meteorology 1% Astrophysics high-risk conjunction events have been developed by space 2% agencies, permitting them to implement avoidance manoeuvres whennecessary[14,17].Non-catalogueddebrisrangingfrom1to 10cmcanalsogenerateverysignificantdamageduetotheirkinetic energy but the collision risk can only be studied statistically Communications throughanalysisofimpactsondedicatedexperimentalplatforms 59% or on launchers and large space debris that return to the Earth Fig.2. Theseveralkindsofactivesatellites.a)Distributionofthe957currentlyactive surface.Table 1 reports collisionprobabilities. These probabilities satellitesaccordingtoorbitclass:LowEarthOrbit(LEO)referstoorbitwithaltitudes areafunctionofparticleflux,which,inturn,dependsonaltitude, between80and1700km;MediumEarthOrbit(MEO)fororbitsbetween1700and vehiclesurfaceareaandtimespentinorbit. 35700km;GeosynchronousOrbit(GEO)fororbitswithaltitudesofapproximately AsimulationmodelofLEOpredictedthat,evenwithnofuture 35700km; Ellipticalorbits haveanon-constant altitude.b)Distribution ofactive satellitesamongsevendisciplinesaccordingtotheiruse.Thesefiguresarebasedon launches,thecriticaltippingpointwherethepopulationofartificial theanalysisofdatafromtheUCSdatabase[13]. spacedebriswouldgrowatafasterratethanthenaturaldecayrate Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 4 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 Table2 Approximateemissions forthe fourmainpropellanttypes(onesolidand three liquid)givenasmassfractionforeachpropellant.Thetotalmassfractionexceeds unitybecauseoftheassumptionthatairmixedintotheplumeoxidizesCOandH2 (source:Rossetal.[23]). Propellanttype N2 CO2þCO H2OþH2 ClOx, HCl Alumina HOx,NOx soot Solid(NH4ClO4/Al) 0.08 0.27 0.48 0.1 0.15 0.33 Cryogenic(LOX/H2O) e e 1.24 0.02 e e Kerosene(LOX/RP-1) e 0.88 0.30 0.02 e 0.05 Hypergolic(UDMH/ 0.29 0.63 0.25 0.02 e Trace N2O4) Localimpactsoflauncheventsaresometimesstudiedbyspace agencies. It is the case for example at the French Guiana Space Center(CSG).Ateachlaunch,about600measurementsaretakenat severaldistancesfromthelaunchzoneandincludeconcentration measurements of hydrochloric acid, nitrogen dioxide, hydrazine andalumina.Thesemeasurementsshowthatimpactsaremainly confinedtothevicinityof thelauncharea(<2.3km)wherehigh levelsofHClandAluminaconcentrationshavebeenrecorded(see Table3). Impactswerefoundtobelowatintermediatedistances(upto Fig.3. RepresentationofspacedebrisinLEO((cid:1)NASA). 8 km) and non-significant beyond. Impacts on water quality, vegetationandfaunaarealsomonitoredanduptonownosignif- icant negative impact has been noted [23]. In order to minimize couldbereachedinabout50years[18].TheeventualitythatLEO localenvironmentalimpactscausedbySpaceactivities,theGuiana couldberenderedinaccessiblebyachainreactionofdebriscolli- sionseperhapsforthousandsofyearseisunderlinedbyseveral Space Centre (CSG) also pays particular attention to both the authors[18e21].Inthiscontext,thesuccessfulChineseFengYun1C transportandstorageofsubstancesforlauncherandsatellitepro- anti-satelliteweapontestin2007,whichsignificantlyincreasedthe pulsion.Fillingproceduresarealsocarefullymonitored;substances which escape during launcherand satellite filling procedures are probabilityofcollision,hasbeenwidelycondemned[18,22]. trapped and neutralized (http://www.cnes-csg.fr/web/CNES-CSG- en[23],). 2.2. Spaceactivityasasourceofpollutioncontributingtothe Besidestransientchangesnearthelaunchsite,whichaffectthe deteriorationoftheEarth’senvironment lowermosttroposphere,emissions,albeitsmall,maycauselasting global changes in the stratosphere. As in the case of aircraft, The risks linked to space activities that are most frequently spacecraftrocketemissionsincludegreenhousegasesthatdirectly discussed in the popular literature are on-orbit collision risks, addtoradiativeforcingandwarming,suchasCO2,andcompounds whichthreatenthecommercialexploitationofspace,andrisksto thatindirectlycontributetoproductionorlossofgreenhousegases peopleonthegroundduringdebrisre-entry.Wewilldiscussnowa suchasozoneandmethane[24].Furthermore,watervapourand topicwhichhasseldombeenaddressed:theroleofspaceactivities soot, which are components of condensation trails, are also asasourceofpollutionfortheEarth’senvironment.Wewillfocus responsibleforpositiveradiativeforcingandcontributetowarm- onenvironmentalimpactsthatarerelatedtolaunch,lifeon-orbit, ing[25].Theamountofemittedgasesistrivialcomparedtoother andend-of-lifestages. sources. For example annual CO2 emissions are estimated to be severalkilotonscomparedtoemissionsofseveralhundredkilotons 2.2.1. Environmentalimpactsduringlaunchandorbitinsertion fromaircraft,which,inturn,accountsforbetween2and3%ofthe Thelaunchstageisresponsiblefortwomainkindsofpollution. totalemissionsfromallactivities[24e26]. The first one is the immediate return-to-Earth of the accelerator stagewhichseparatesfromthelauncherafterfuelexhaustion.The accelerator stages are not systematically salvaged and seldom Table3 reused.Togiveanideaofthematerialquantitythatcanreturnto ExampleofmaximalconcentrationsofHCLandaluminameasuredduringanAriane Earth,thetwoemptyacceleratorsofAriane5(mainlycomposedof 5launch(flight185,August24,2008).Nearfieldreferstoadistancefromlaunchsite steel)weighabout38tonseach.Thesecondsourceofpollutionis <2.4kmandfarfieldfrom2.4e24km.Measuresarecomparedtohumantoxicity thresholds(source:http://www.ggm.drire.gouv.fr/). relatedtopropulsionsystemfunctioning.Mostspacecraftsdepend onarocketengineforpropulsion.Approximateemissionlevelsfor Maximalnear Maximal Toxiclimitsdefined themainpropellanttypesaregiveninTable2. field farfield forhumans concentration concentration (mg/m2) (mg/m2) Table1 IonCLe 5136.2 89.84 90mg/m3:irreversible Probabilityofcollisionoveroneyearaccordingtodebrissizeforasatellitewitha (HCL) (measured effectafter30min 20m2surfaceareaatanorbitalaltitudesimilartoSPOT,i.e.825km.(Source:CNES at4.35km) exposure,700mg/m3: [16]). lethaleffectafter30min exposure Debrissize >0.1mm >1mm >1cm >10cm Alumina 94.68 3.49 Acceptablemeanexposure Probabilityofcollision 1 0.5 3(cid:2)10(cid:3)3 2(cid:2)10(cid:3)4 valueforworkers¼10mg/m3 over1year during8h,5days/week Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 5 However, rocket combustion products are the only human- characteristics of several polymers. The experimental data they produced pollutants injected directly into the middle and upper usedshowsthaterosionyields(expressedasthevolumelostper stratosphere. Up to now, few authors have studied phenomena incident atomic oxygen atom in cm3/atom) vary from a factor of occurring in higher atmospheric strata associated with rocket about90betweenthemostandleastresistantpolymers.According emissions.Impactsofemissionsinthestratosphereareliabletobe totheseresultsKapton,acommonlyusedspacecraftmaterial,isa more important than impacts of emissions occurring in the moderatelyresistantpolymer.ImportantKaptonmasslosseshave troposphere. First, atmospheric circulation in the stratosphere is alreadybeenobserved,e.g.upto35%reportedbyNASAontheSTS3 characterizedbyfasterhorizontalmixingofgases.Thusemissions shuttle mission [29]. Even if this announced rate is open to dis- will spread throughout the stratosphere layer and will be longer cussion due topossible inconsistency in dehydratation states be- lasting [24]. Second, the specific composition of the stratosphere tween pre- and post-flight mass measurements [28] the above will give rise to specific reactions. The stratosphere includes the mentioned studies demonstrate that a significant portion of the majorpartoftheozonelayerandischaracterizedbyalowwater insulationmaterials usedtoprotectthe satellites canbereleased vapour concentrations. Indeed, atmospheric water vapour con- intotheLEOdomainintheformofvolatileoxidationproducts. centration decreases with altitude and over 99% of water vapour lies within the troposphere. While climate response seems to be 2.2.3. Environmentalimpactsassociatedwithsatelliteend-of-life independent of where CO2 emissions occur (http://www. The on-orbit lifetime of non-active satellites and other debris co2offsetresearch.org/aviation/DirectEmissions.html),the increase dependsonthepresenceanddensityoftheterrestrialatmosphere. in forcing due to water vapour emissions in the stratosphere is Atmospheric density decreases the greater the distance from the significant compared to a similar water vapour emission in the Earthbut,atagivenaltitude,italsovariesasafunctionofseveral troposphere [25]. Emitted compounds also contribute to ozone factors including solar activity and latitude. Atmospheric drag depletion in several ways. Some of these substances are highly slowsdownorbitingobjects,leadingtotheirreturntoEarthwithin reactive radicals e NOx, HOx, ClOx e that are directly involved in atimeperiodthatdependsontheorbitaltitude(e.g.ISSlifetime catalytic cycles, thusleadingtoan increasein theozoneremoval orbiting at 300e400 km would be w6 months to 1 year, SPOT rate.Ascatalysts,theycanhavehugeimpactsevenifpresentinonly lifetimeorbitingat825kmeabout200years)[16].Duringatmo- trace amounts. A single radical molecule can destroy up to 105 spheric re-entry objects are intensively heated and part of the ozonemoleculesbeforebeingdeactivatedeforexamplethrough material is sublimated, thus slightly changing atmosphere reactionsthatremoveradicalsfromthecyclebyformingreservoir composition, though probably with insignificant impact. Large species eand transportedoutof the stratosphere.Otheremitted piecesofdebriscanalsoreturntoEarth.Onaverageapieceoflarge compounds contribute to an increase in tropospheric radical res- orbitaldebris(radarcross-section>1m2)fallsbacktoEarthonce ervoirs.Forexample,particlessuchasaluminaand,possiblysoot or twice a week. Most of them (>70%) will impact water bodies particles, are responsible for the liberation of radicals from the [30]. Casualty risks are more closely associated to natural re- radical reservoirs present in the stratosphere. Concerning solid entries,includingprematureonesasinthecaseoflaunchfailure. rocketmotors,emittedHClisinitselfaradicalreservoir.Further- LaunchfailureratesreportedfortheApril2009eSeptember2009 more,whilewater vapouremissions are widelyconsidered inert, periodwasonefailureforevery39launches[31].Casualtyrisksare H2O, whichisemittedbyallrocketengines,is thesourcegas for currently estimated to be lower than the risk associated with HOxradicalsandcontributestotheformationoficeparticlesalso meteorite impacts [16]. However both controlled and natural re- responsible for ozone loss. Ozone loss linked to water vapour is entriesareapotentialsourceoforbitaldebrisreturningtoEarth. highly nonlinear and difficult to predict [24]. The ozone layer is For instance, during the controlled Mir re-entry, while the initial protectedbyinternationalagreementsthatlimittheproductionof massinorbitwasaround140tons,30tonsofdebrisfellintothe substancescausingozonedepletion(i.e.theMontrealProtocolon PacificOcean[16]. Substances That Deplete the Ozone Layer). Ross et al. [24] There-entryofPhobos-Gruntisarecentexampleofaccidental demonstrated that if the Space Shuttle had met its original goal re-entry.ThemainobjectiveofthisRussianmissionwastostudy of weekly launches it would have been responsible for an ozone Phobos, one of the moons of Mars. The satellite was launched in lossclosetoaquantitythatRossetal.[24]deemedtobetheupper November 2011 but the spacecraft failed to exit the Earth’s orbit limit acceptable to the international community that established andfellintothePacificOceanonJanuary15,2012.Onboard,there theMontrealprotocoltoprotectthestratosphericozonelayer,even remained about 11 tons of unused highly toxic propellant: the whentakingintoaccounttheuniquecontributionofspaceactivity. unsymmetricaldimethylhydrazine(UDMH).Fueltanksarelikelyto haveexplodedhighabove the Earthandsome experts suggested 2.2.2. Environmentalimpactsduringon-orbitlife that the propellant was burnt. However other experts think that Severalcompoundsarealsoreleasedintheupperatmospheric partof thematerial thatvaporizedcouldhavere-condensedinto layers during the on-orbit lifetime of LEO satellites. First, the at- smallparticlesthatmayremainintheupperatmosphereformany mosphericdraginLEOcausesorbitaldecayandtheplatformhasto yearsandinfluence,evenifinaminorway,atmosphericchemistry be repositioned occasionally. This is usually performed using (http://www.space.com). nozzle-basedsystems,andhydrazineisthemostfavouredmono- There-entryoftheUpperAtmosphereResearchSatellite(UARS) propellant. The highly exothermic catalytic decomposition of hy- in September 2011 is another recent example of uncontrolled drazineproducesjetsofhotgasandthusthrust.Theemittedgasis natural re-entry, but one which occurred after the satellite had composed of ammonia (NH3), hydrogen (H2) and nitrogen (N2). ceased its scientific life in 2005. NASA estimated that twenty-six Second,thepresenceofadiffuseatmosphereslowlyerodessatellite satellite components, weighing a totalof morethan 500 kg, may platforms.Atomicoxygen,thepredominantcomponentintheLEO havereachedtheEarthsurface(http://www.nasa.gov). atmosphere, is responsible for the degradation of thermal, me- Anotherenvironmentalthreatfromspaceactivitiescomesfrom chanical, and optical properties of exposed materials [27]. It in- the use of nuclear reactors. Such reactors generate substantial teracts with hydrocarbon polymers (e.g. Kapton, Teflon, Mylar.) amountsofelectricalpower.Theiruseonmilitarysatellitesisbehind thatareusedtothermallyinsulateandprotectpartsofthesatellite. theincreasedspatialresolvingpowerofon-boardradarsanddoes In a recent study, Banks et al. (2011) [28] developed a model to away with the need for large solar sails. Decreasing the satellite assess the oxygen erosion yield according to the molecular cross-sectionalareaisparamount,therebymakinglocalizationmore Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 6 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 difficult and lowering risks of hostile actions from anti-satellite 3.1. Integratingenvironmentalconsiderationswhendesigning systems (ASAT) [32]. The radioisotope thermoelectric generators systems have also enabled missions such as the Apollo Lunar surface ex- perimentsandinterplanetarymissionsthatrequiretraveltoareas Firstweshouldshiftfromaviewwheresystemsaredesignedto wheresunlightintensityandtemperaturesarelowandwherera- meet measurement objectives that consider primarily economic diationbeltsareverysevere[33].Since1961,theUnitedStatesand andtechnicalconstraints,toamoreholisticviewthatconsidersan theformerSovietUnionhaveflownatleast43radioisotopether- individualsystemasacontributortothewholesatellitesetandthat moelectricgeneratorsand36nuclearreactorstoprovidepowerfor includes consideration of interactions between remote sensing respectively24US and37 Soviet singleor multiplereactor space systems and the Earth’s environment, including environmental systems[33].Severalnuclear-poweredspacevehiclesareknownto impacts from the system production stage to its end of life. This havefallentoEarth,e.g.Transitin1964orCosmos-954in1978,and would enable the design of measurement systems capable of wereresponsibleforthereleaseofradioactiveelementsintheat- providingvaluableinformationforsustainablemanagementofthe mosphereandontheEarth’ssurface[32].Currentknowledgemakes Earth resources while reducing to a minimum their inevitable itverydifficulttoassessenvironmentalimpactsandtheamountof environmental impact. This would thus help remove the contra- radioactivity that would reach the Earth surface in case of the dictionthatcurrentlyexistsbetweentheenvironmentalimpactof disintegration of a reactor core, in particular in the upper atmo- space activities and the purpose of the Earth remote sensing sphere,duringaccidentalre-entry[32]. mission,whichistoassistsustainablemanagementoftheplanet. InSection2.2,wehaveonlydiscussedenvironmentalimpacts Butthiswouldalsorequiremajorandchallengingchanges. fromlaunchtosatelliteend-of-life.Duringlaunchesandre-entries Acquisition processes are currently designed based on usual chemical components and debris are released in all atmospheric measurement qualitycriteria, e.g. degrees of precision and of ac- layers and part of them falls back to Earth. Based on current curacy,andaremainlydeterminedbyeconomiccosts,technolog- knowledgewecanonlysaythat,duetoitsspecificcharacteristics, icalreadinesslevels,andtechnicalconstraints.Aperspectivethat spaceactivityisasourceofcasualtyriskandalsocontributestoEarth considers sustainability would also include environmental and pollution with a pollution capacity difficult to assess but with socialdimensions.Astheenvironmentunderpinsbothsocietyand potentiallyhighrisks.Butformanymanufacturedproductsasig- the economy [36], an essential step in designing measurement nificantportionoftheenvironmentalimpactisnotdirectlylinkedto processes that respect sustainability principles is to integrate product use but rather to its manufacture and transport [34]. environmental dimension when evaluating the data acquisition Therefore,impactsassociatedwithlauncher,platformandinstru- process as regards sensor performance and data quality. For mentmanufacturing,andthoseimpactsrelatedtothefunctioningof instance,shouldacountryspend$1billiononitsownspaceradar the ground segment should also be carefullyexamined. Unfortu- toensurecontrolofthedataifbetterorsimilardatacouldbeac- natelytheseimpactshavenotbeenstudieduntilnowandwefound quired less expensively by (1) cooperating with international nostudyreportorpaperdealingwiththistopic.Theonlyactionswe partnersor(2)utilizingaircraftradarsystemstobettereffect?The couldidentifyinordertofillthisgaparethoserecentlyinitiatedby correct answer to this question becomes even more obvious if theEuropeanSpaceAgency,ESA,withintheframeoftheCleanSpace environmental concerns are taken into account, though such initiative[35].Firstresultsshouldbeavailablesoon. treatment excludes more ambiguous though important concerns suchasnationalprideandmaintenanceoftechnologicalreadiness. Toolsdesignedtoassessenvironmentalimpactsexistandcould 3. TowardsustainableEarthobservationsystems beusedatseverallevelstobettertakeintoaccountenvironmental concernsintheevaluationofEarthobservationdataandproducts Remotesensingscientistsviewsatellitedataacquisitionasthe supplychain.LePellec-Dairon[37]proposedanovelapproachto first stepin an information stream that ultimatelyleads toinfor- better assess the environmental value of Earth observation data. mative products, decisions and actions based on those products, This approach requires one to evaluate the economic value of additionalquestionsstemmingfromrecentresearch,andperhaps environmentalgoodsandtomovefromthisvaluetothevalueof follow-on missions to address the most important outstanding Earth observation data that are used to improve Earth resource questions.However,priortotheacquisitionofthefirstbyteofdata, management. Certainly, assigning an economic value to an envi- considerationshouldbegiven,touncertaintiesassociatedwiththe ronmentalgoodcanbechallenging,possiblysubjectiveandprob- durabilityofEarthobservationfromspaceandalackofknowledge ablyarguable.Aswithanyqualitativeversuseconomiccomparison, ontheconsequencesofspaceactivitiesontheEarth’senvironment. onecentralissue becomesassigningadefensible monetaryvalue Also, in keeping with the transparency1 quality of a measuring to,forinstance,avista,ahikingpath,acleanriver,or,inthecaseof instrument,asacommunity,weshoulddoourbesttoensurethat satellite missions, closure of the global carbon budget, a 10% in- themethodsusedtoobserveandmeasuretheEarth’senvironment creaseintheprecisionofa biomassestimate, oramoreaccurate interfere least with that environment. For example, when mea- map of deforestation. That said, Le Pellec-Dairon [37] developed surementsareacquiredtoassesscarbonpoolsandfluxes,withthe suchanapproachtohelpimprovetheefficiencyofthestrategyfor aimofbetterunderstandingthecarboncycleandclimatechange, Earthobservationmissionmanagementandtoselectthemissions themeasuringsystemshouldcontributeaslittleaspossibletothe with the largest positive effects, including economical, political, globalcarboncycle. technical and environmental values. However, the proposed We suggest points of view that, if shared, naturally lead to approachonlyconsidersthetotaleconomicvalueoftheenviron- considerationofquestionsconcerningtheenvironmentalsustain- mentalinformationsuppliedbyEarthresourcesmissionsandwill abilityofEarthobservationsystems. remain incomplete while this value is not weighed against the economic value of environmental degradation due to the manu- facture,launch,on-orbitmaintenance,andeventualdestructionof 1 InthefieldofMetrologyandUnitsofMeasure,transparencyistheabilityofa themissionhardware. measuringinstrumentnottoalterthemeasurandetermanddefinitionapproved WethinkthatusingtoolssuchasenvironmentalLCAcanbean byISO,theInternationalBureauofWeightsandMeasures(BIPM),theInternational effectivewaytoassessenvironmentalimpactsofspaceactivityand Electrotechnical Commission (IEC), the International Organization of Legal Metrology(OIML). to improve sustainability of Earth observation systems and thus Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 7 increase efficiency and environmental value of the Earth data error,i.e.reducingthechanceofacceptingthenullhypothesiswhenit production chain. Environmental LCA methods and procedures isfalse[9].UsualscientificapproachestendtominimizeType-Ierror. havealreadybeenstandardized(ISO14040andISO14044).LCAare Thisallowsustoachievehighlevelsofconfidenceinthedecisionin multi-criteria, quantitative approaches that enable assessment of ordertorejectanullhypothesisandacceptthatsomesortofchange environmentalimpactsassociatedwithallthestagesofaproduct’s hasoccurredorthata“new”conditionexists.Whenappliedtoenvi- lifefrom-cradle-to-grave(i.e.fromrawmaterialextractionthrough ronmentalmanagement,minimizingTypeIerrormeansthatweneed materials processing, manufacture, distribution, use, repair and tobevirtually-certainthatenvironmentalorecologicaldamagehas maintenance, and disposal or recycling) [34]. Environmental im- occurred(duetospaceactivityinourcase)beforewewouldacceptthe pactsareassessedthrougheithermidpointorendpointindicators alternative hypothesis [9]. The net effect of relying on such an derived from the inventory of fluxes of both raw materials and approachewhendeterminingifandwhatactionsmustbetakeneis pollutants. Ozone depletion, climate change (in practice all the thatthedamagewouldhavealreadybeendonebythetimethetest greenhousesgazesemissionsaretransformedintoanequivalentof alertsusoftheneedtoact.Withrespecttoprevention,mitigation,or emitted CO2 according totheirwarming capacityand taking into remediation,thiscourseofactioniscounterproductive.Inorderto accounttheirlifetime),landusechange,acidification,nutriphica- calculateTypeIIerror,theanalystwouldhavetodefineanalternative tion,naturalresourcesdepletion(suchasminerals,fossilfuel.)are hypothesis, i.e. that condition beyond which intolerable damage examplesofindicatorsprovidedbyLCA,whenamidpointimpact wouldbedoneoratippingpointisreached.Suchanalternativecould category, or problem-oriented approach, is chosen. Impacts are taketheformofachemicalconcentrationataparticularaltitude,a translated into environmental themes. Endpoint modelling, also radiation level, or a collision probability. Furthermore, for low- calleddamage-orientedapproach,canalsobeusedbutisconsid- probabilityhigh-impacteventsee.g.LEObecomesinaccessiblefor ered more uncertain. Indeed endpoint impacts result from the hundredsofyearseTalebetal.[38],whoanalysedcommonerrorsin combination and the transformation of midpoint environmental thewayrisksaremanaged,maintainthatitismoreefficienttoreduce impactsintoissuesofconcernsuchasdamagesonhumanhealth theimpactofthreatswecannotcontrolratherthantofocusonsta- (e.g. numberof cancers), on ecosystems (e.g. loss of habitats), on tisticalpredictionsofsuchevents.Thisiswhy,thedevelopmentof naturalresources. spacedebrismitigationandremovaltechnologieslikethoseconsid- We suggest that environmental LCA methods and mid-point eredbyESAinthecontextofitsCleanSpaceprogramareessential. impact indicators be considered as one of a number of tools and Debrismitigation/removalresearchcomplementsresearchconducted metricsthatcouldbeemployedtofacilitate,forinstance,compar- tocollectinformationonsmalldebris,i.e.thoseobjectsnotdetected isons between different mission scenarios or to optimize a given withcurrentmethods,andtomodelthebehaviourofthespacedebris mission.Inadditiontoimprovingthecoherencebetweenthedata environment[35].Innovationinthisareacanalsohavespin-offben- acquisitionstepandtheendgoalsofsomespacemissions,itcanbe efits forthe spaceindustrysectorallowinganearlypositioningon arguedthatactingtodevelopaspaceactivitywithreducedenvi- emergingmarketsthatmightreplacepartofthespacesectortradi- ronmental impacts on both Earth and space has advantages for tionalmarkets if a stronger regulation of the use of space became spacesector.ESAhasadoptedapioneerpositioninthisareawithits effective. CleanSpaceinitiative[35]and,amongtheargumentsputforward tojustifythisinitiative,someareworthhighlighting: 3.3. Revisitingcommonideasintheremotesensingandspace activitysectors - Laws evolve, and the space community should expect envi- ronmental law to eventually address space-related activities, Itwould be useful toqualify three commonly heldarguments especiallythosehavingtodowithlaunchandspacedebris.The for remote sensing data that are taken as a given. First, from a development of eco-friendly space policies and programs is strictlyeconomicalpointofview,spaceremotesensingisconsid- onewaytoproactivelyaddresslegalevolutionsthatareliable eredcosteffectiveforend-users,especiallygiventhatmuchofthe toleadtodisruptionofclassicalspace-qualifiedmaterialsand data are freely available, e.g. Landsat, MODIS. But this statement processes. Among candidate green technologies are develop- does not take intoaccount all national funding invested in space mentofgreenpropellantstoreplacehydrazineorothertoxic activity. Consideration of the full range of expenditures might liquidpropellants,replacementofchromatesusedforsurface indicatethat,forcertainspacemissions,airbornesolutionsmaybe coatingandbonding,greenelectronics,andreductionsinbulk more viable and may provide better data from the standpoint of machiningthataresourcesofrawmaterial waste.Hydrazine seasonality, spatial coverage, and higher radiometric and spatial hasalreadybeenincludedinthelistofsubstancesofveryhigh quality at a lower cost. This is especially true when the mission concernbytheEuropeanCommunityRegulationonchemicals objective addresses national or regional concerns rather than andtheirsafeusethatdealswiththeRegistration,Evaluation, continental or global concerns. Second, field data are often AuthorizationandRestrictionofChemicalsubstances(REACH). described as being costly and time consuming [39] and in many - Byactingproactively,ESAexpectstobewellpositionedtohelp instancesthiscommonideaiscorrect.Insouth-easternAlaska,for shape future laws, thereby making it easier to comply with instance, the cost of establishing and measuring one US Forest futureregulations. Servicefieldplotin2013isestimatedtobe$9000USD.Thatsaid, - Theactofencouragingindustrialiststodevelopandadopteco- suchcostsshouldnotimplythatremotesensingdata,sometimes designsandeco-technologiespromotesinnovationandisthus “free”, can replace field observations. Field measurements are anefficientwaytomakeindustrymorecompetitiveinworld essential for calibration/validation steps in most remote sensing markets. data processing approaches, and also provide information that cannotbeacquiredbyothermeans,e.g.assessmentoflocalbiodi- versity,soilproperties,presence/absenceofrarespecies,estimates 3.2. Thebenefitsofchangingourrelationshipwithstatistics ofdownedwoodydebris,regeneration.Remotesensingwillnever replace the need to make field measurements and maintaining Secondly, with regards to environmental risk management, it sufficient field observation networks and field expertise remains wouldbenecessarytoshiftfromminimizingType-Ierror,i.e.rejecting essential.Third,itisworthnotingthat,forthespaceactivitysector, thenullhypothesis(orstatus-quo)whenitistrue,toreducingType-II anyobject(satellite,pieceoflauncher.)isconsideredasrecycled Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 8 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 whenithasbeendestroyedduringitsre-entryintotheatmosphere - a spatial sampling in order to have a mean canopy height orhasfallenbacktotheEarth’ssurface.Thisisveryremotefromthe within1e2mfor1haand1kmgridcells(50cloud-freeob- notionofrecyclinginthecontextofsustainabledevelopment.The servationsper250mgridcellwouldberequiredtoachievethe Earth’s capacity to sequester human waste is limited [36] and requiredheightaccuracyat250m); recycling aims to reduce waste. Furthermore a part of so-called - contiguouslysampledprofilesforprimarilyecologicalreasons, “recycled”spacecraftisinrealitycomposedofpollutantsemitted e.g. estimation of height correlation length scales and height intotheatmosphereordepositedontheground. distributions,butalsotechnicalandpracticalones,e.g.ground findingandlocalstructuralcontext; - atemporalresolution enabling theproductionof twoannual 4. Potentialmeasurestoimproveenvironmental biomass maps in order to capture seasonal variabilityand to sustainabilityofEarthobservationsystems monitor annual biomass change and a minimum of a 5-year observation period to capture the successional dynamics of Embracing the unconventional points of view presented in the forestecosystems. previoussectionleadstosuggestionsonsuitableactionsthatcould helptodesignmoreefficientEarthobservationsystemsandnetworks. ICESat,anicemissionwhichcollecteddatausefultoterrestrial Inordertoillustratesomeofthesuggestedactionswewillrely studies from 2003 to late 2009, was the first spaceborne lidar onastudycase:thedesignoflidarsystemsforforestmonitoring. system designed to measure terrestrial surfaces. Despite the ice- The first sub-section presents this study case and sets out the centric design, several studies used ICESat/GLAS data to estimate generalcontext,therequireddataonsustainablemanagementof foreststructureandbiomassatregionalscales[56,59e64].Lefsky forests,andsomespacebornesolutionsthathavebeenunderstudy. [64],Simardetal.[65],andLosetal.[66]havedemonstratedthe The second subsection suggests some measures that might be global capabilities of space-based GLAS measurements while adoptedwhenplanninganEarthobservationspacemission.Inthe Nelson[67]outlinessomeofthelimitationsassociatedwithGLAS- thirdsubsectionthescopeofthediscussionisextendedtoincor- based biomass estimates. ICESat-2, scheduled for launch in 2016, porateairbornemissionsinanEarthobservationnetwork. willbethesecondlidarmissiondedicatedtoEarthsurfacemoni- toring. It is also primarily designed to characterize and monitor 4.1. Studycase:lidar,avaluabletechnologytoimprovesustainable polar ice, and includes among its secondary scientific objectives managementofforestecosystems measurementofvegetationcanopyheightasabasisforestimating large-scalebiomassandbiomasschange[68].Howeverdatasim- While sustainable management of forests is recognized to be ulationshavedemonstratedthatICESat-2,initscurrentdesign,will crucial for the future of mankind, forest ecosystems are heavily notreplacetherecentlyshelvedDESDynIvegetationlidarmission impacted by climate change, natural hazards and anthropogenic [69]. Even under ideal conditions e clear sky, night-time data pressures,particularlydeforestationanddegradation.Overthelast collect,minimaltopographyeweexpectthegroundsignaltobe decade, about 13 million hectares (0.33% of total forested area) lostatcanopyclosuresexceedingw96%,thusmakingcalculationof wereconvertedtootherusesorlostthroughnaturalcauseseach canopy height problematic in closed-canopy forest growing in year[40].Robustsystemsformeasuring,assessing,andreporting areaswithappreciablelocaltopography. keyforestparameters,e.g.biomass,carbon,mustbedevelopedin Scientistshavedesignedandproposedtospaceagenciesspace ordertodefineappropriatemanagementpracticesandpoliciesto lidarmissionswithforestmeasurementandmonitoringasprimary address the challenge of sustainable management of forest re- scientific objectives (e.g. VCL, LVTM, Carbon-3D, DESDynI-Lidar, sourcesandtostrengthenforest-basedclimatechangemitigation LEAF,SpeCL.[69e74]).AllproposalseUS,ESA,andJapanehave strategies [41e44]. The utility of lidar has been widely demon- thus far been unsuccessful; none havebeen launched. The Vege- stratedwithrespecttoforeststructuremeasurementsandbiomass tationCanopyLidar(VCL)wasselectedinMarch1997byNASAas estimation[45e54].Thisisalsothecaseinclosed-canopytropical thefirstEarthSystemSciencePathfinder(ESSP)spaceflightmission areas supporting high biomass forests greater than 200 t ha(cid:3)1 and was scheduled for launch in January 2000 [75]. The mission [55,56], where optical vegetation indices and volumetric radar was discontinued due to unexpected technical limitations that measurementstypicallysaturate[57]. precludedconstructionofthelaserswithintheinitialquotedcost. Consequentlyaspacebornelidarthatacquiresvegetationheight Thereasonsunderlyingthedecisiontodiscontinuethelidarpartof and canopy closure measurements, when employed as a global the DESDynI mission, avegetation measurement and monitoring sampling tool e either alone or in combination with optical and missioninitiallyapprovedbyNASA,areunknowneventoscientists radar imagery e and supported by ground observations would thatweredirectlyinvolvedintheproject.ItwascancelledbytheUS appeartobeapromisingsolutiontoestimateabovegroundforest governmentasanausteritymeasure,butweguessthatitmayhave biomassandcarbonataglobalscale.Indeedsuchasolutionwould been jeopardized due to what was seen by budget or science combine the beneficial measurement properties of spaceborne managersasmissionoverlapwithICESat-2.WithrespecttoESA’s remote sensing, i.e. acquisition of global information that is LEAFproject,theevaluatorsestimatedthatthemissionwastech- consistentbothinspaceandtime,withlidartechnology.Adetailed nicallyfeasiblebutabovebudget(þ15%abovetherecommended review of measurement requirements for the assessment of industrial ceiling cost). The evaluators acknowledged that the biomass,biomasschangeandcarbonflux,biodiversityandhabitat, missioncouldhavebeencomplementarytotheBIOMASSmission, is provided by Hall et al. [58]. The authors identified several re- aP-bandSyntheticApertureRadar(SAR)alsodedicatedtobiomass quirements for lidar measurements which notably include the estimationthatwasselectedasoneofthethreecandidatesforthe following[58]: 7th Earth Explorer mission. The evaluation team also mentioned the possible mission overlap between LEAF and ICESat2 or DES- - globalcoverageofforestedareas; DynI,whichwasnotyetdiscontinuedwhentheLEAFprojectwas - basedon25mfootprints,measurements,offorestheightwith submitted.Thechosenrevisitperiod,i.e.3years,andthefactthat 1maccuracyatzeroslope,offorestverticalstructurewitha1e thesameindividuallaserpulselocationscannotberevisitedwere 2maccuracyandofcanopycovermeasuresaccuratetowithin identified as shortcomings. SpeCL, a large footprint multi- 10%; frequency lidar with the capacity to revisit previous spots, was Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 9 notselectedbutratherwasshortlistedbyESA.Thismaybeinter- couldcontributetotheestablishmentofsuchinternationalpolicies pretedasanadditionalevidenceoftheacknowledgementbyESAof [15].HowevertheconsequencesofheedingtheadvicefromTaleb the scientific interest of vegetation space lidar missions, even if etal.[38]concerninglow-probabilityhigh-impacteventswouldbe from a technological point of view revisiting previous spots is, 1) todo everything in our power toreduce space junk by devel- practicallyspeaking,currentlynotachievableinthecivilianspace oping,forexample,orbitaldebrisremovaloperationsasproposed sector. by Weeden [20], 2) to do our utmost to mitigate future launch InMay2013BIOMASSradarmissionwasselectedbytheESA’s pollution and debris, therebykeeping space as pristine as practi- EarthObservationProgrammeBoardtobecometheseventhEarth cally possible, and 3) to think about alternatives to spaceborne Explorermission.However,whilewehavethebenefitofhindsight solutionsincaseLEObecomesinaccessible. aboutthecapacityoflidarsystemstoprovidevaluableinformation Even if it has a limited contribution to current space debris forbiomassestimation,therearestilluncertaintiesandopensci- environment,ESAclearlyexpressesitswillingtotacklequestions entific questions about the potential of the BIOMASS system to concerning space debris mitigation and remediation within the providetheexpectedinformation,inparticularinforestswithvery frameofitsCleanSpaceprogram.Amongtheplannedactionsare high biomass levels. In 2011, several projects were submitted to thefollowing[35]: NASA(GEDIproject),JAXA(i-LOVEproject),andESA(GRAILproj- ect) that would involve embedding on the International Space - for space-debris mitigation: develop compact, robust and Station (ISS) eitheravegetation lidar system ora blue and green autonomoussystemsforde-orbitingofspacecraftinLEO,de- carbonsystem,i.e.adualfrequencylidartoassessbiomassbothon velopend-of-lifepassivationofpropulsiveandpowersystems land and in oceans and coastal waters. These projects would to lower the risk of spacecraft break-up, develop spacecraft partiallyreplacetheDESDynIlidarmission.Onlytheprojectsub- designedtobreak-upduringre-entryinsuchawayastoavoid mittedtotheJAXA,renamedMOLI,iscurrentlyselectedforfurther large space debris impacting the Earth surface (“Design for studies. The project submitted to NASA was rejected despite the Demise”concept); factthattheevaluationcommitteeacknowledgeditshighscientific - forspace-debrisremediation:develophardwaretoapproach, value[76].InDecember2012,thedecisionofnon-selectionforthe dock with,andcontrol aspacecraftscheduledforde-orbitor GRAILprojectwasalsocommunicatedbytheESAtotheprincipal sizeabledebrissoontode-orbitsoastofullycontrolreentryto investigator. assurethebestenvironmental(andsocietal)outcome. Considering this international context, we will rely on this exampletoillustratesomeoftheactionsthataresuggestedinthe 4.2.2. Mitigationofenvironmentalimpacts followingsubsectionsinordertohelptoincreasesustainabilityof Tobettergraspthisissueofdamagerelatedtospaceactivityon Earth observation systems, with a focus on systems dedicated to theEarth’senvironment,researchisneededtoassessthebehaviour forestbiomassmonitoring. ofmaterialsreleasedintotheupperatmosphericlayersbylaunch vehicles,fuels,andre-entrydebris.Auniquecharacteristicofspace 4.2. Improvingsustainabilityofspaceremotesensing activityisthatitistheonlyhumanactivitythatreleaseselements intotheupperatmosphericlayerswhereconcentrationsofnatural Duetogeopoliticalconstraints,satellite-basedsolutionsseemto compounds are low. Consequently even the introduction of ele- betheonlywaythatiscurrentlyavailabletoacquiredataglobally. mentsinsmallquantitiescangreatlyaffectatmosphericcomposi- Furthermore theyguarantee ahigherdata consistency thanwhat tionandchemistry;asevidencedbytheeffectsofreactiveradicals can be expected from airborne solutions that are managed at on the ozone cycle [24]. Impacts of space activity on complex regional or national levels and also facilitate user access to data. processes occurring in upper atmospheric layers have been little However, when designing space missions, actions that are in studieduntilnowandthereisaclearlyalackofknowledgeonthe keepingwiththeabove-mentionedconsiderationscouldbetaken, potentialconsequencesofspaceactivityonatmospherecomposi- thereby making them more compatible with environmental sus- tionandonradiativetransfer. tainabilityprinciples.Twomainchallengeshamperthesustainable Inadditiontositeimpactstudiesperformedonindustrialsites development of space activity: space debris, which threaten the (e.g.launchsite),environmentallifecycleassessment(LCA)could activityitself,andthesector’spotentialnegativeimpactsonboth beusedmoregenerallytohelpquantifytheenvironmentalimpacts spaceandEarthenvironments. ofEarthobservationsystemsandtoidentitycriticalstageswhere theseimpactsmightbemitigated.EnvironmentalLCAcouldalsobe 4.2.1. Mitigationandremediationofspacedebris usedtocompareseveralmissionscenariosandtohelpchoosethe AccordingtoWilliamson[77],ethicalandcodepolicyforspace scenario that will provide data meeting the information re- shouldincludeprotectionofthevariousorbitalzonessurrounding quirementsspecifiedbyscientistsandend-userswhileleadingto the Earth given their importance as a commercial and scientific thelowestenvironmentalimpactlevel. resource by formalizing debris mitigation measures. Space Inthisfield,ESAhasadoptedapioneerpositionandhasstarted agencies,e.g.CNES,ESA,NASA,havealreadydevelopedguidelines usingLCAtostudyenvironmentalimpactsofawholemission,from tomitigatespacedebris.Ataninternationallevel,theUnitedNa- theearlyresearchstagestothemissionend-of-life,includingallthe tionsCommitteeforthePeacefulUsesofOuterSpace(UNCOPUOS) tests required to check the system resistance to hostile space hasalreadytakenakeeninterestinspacedebrisandintheuseof environment, i.e. high temperature gradients and radiation. nuclearenergyinspace.But,asaconsultativebody,ithasnoleg- Convinced of the benefits of green-technology development (see islative power [21,77]. Bradley and Wein [18] demonstrated that Section3.1),oneofESA’sobjectivesistogeneralizetheassessment achieving full compliance with the 25-year spacecraft deorbiting of environmental impacts of space activity and to develop a guidelinescouldmaintainthelifetimeriskfromspacedebrisata framework,withdatabasesandtools,dedicatedtothisactivitythat sustainable level. Williamson [15] underlines that despite their could be used by European space agencies and industry [35]. A potential efficacy, technological solutions to space sustainability preliminaryversionofthisLCAtooliscurrentlybeingimplemented will not be effective without the development of international to model environmental impacts of missions (launchers and sat- policies and laws to support and promulgate them. He presents ellites)representativeofvariousapplicationfields[35].Firstresults severalinitiatives and placesfordiscussion andnegotiationsthat willbeofgreatinterestasitisthefirststudyinthisfield. Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003 10 S.Durrieu,R.F.Nelson/SpacePolicyxxx(2013)1e13 Furthermore,ifweseeeachmissionasanactivitycontributing vegetation, though dense, may be convolved with the pulse tospaceactivityasawholeandthereforeasacontributortothe width (ICESat/GLAS, 6 ns FWHM) such that waveforms from problems discussed in this paper, very simple measures can be flat surfaces and waveforms from shrublands are essentially proposedtoimprovesustainabilityofspaceactivityfromthispoint identical. ICESat-2, a photon-counting, moderate footprint forward.First,ifweconsiderspaceasalimitedresourceweshould (10 m diameter) space lidar due for launch in 2016, may be strivetouseitparsimoniouslyandtolaunchasatelliteembeddinga unable to retrieve ground reliably under very dense forest measurementsystemonlywhenahighlevelofconfidenceonthe canopiesthatexceed95e96%canopyclosure.Tobefair,ICESat- system’s capacity to meet measurement requirements has been 2isanice missionthatisexpectedtomeetstringent perfor- reached. Second priority might be given to missions with longer mancecriteria associated with measurementand monitoring lifetimesinordertoreducethenumberoflaunches,spacedebris, of ice sheets, sea ice, and glaciers. Vegetation measurements andtolessenenvironmentaldamage.Third,prioritymightalsobe areofsecondaryconcern. given tointernational missions in ordertoreducemission dupli- - Evaluate the added-value created by coupling a very high cation,e.g.multipleX,C,L-bandradars,multiplew30mLandsat- spatialresolutionmultispectralimagerwiththelidar.Suchan like clones, or multiple satellite constellations embedding very imager, if accurately boresighted, would provide contextual highresolutionimagers.Considerationshouldalsobegiventothe andgeolocationinformationcontemporaneouswiththelidar useofexistingspaceinfrastructure.Forinstance,theInternational acquisitionthatcouldhelptoimprovebothaccuracyofinver- SpaceStation(ISS)hostsrelativelyshort-lived(e.g.2year)systems sionmodelsusedtoestimateforestparametersfromlidardata thatallowinstrumentengineerstotest,refine,andimprovetheir andextrapolationofbiomassestimations.Thiscouplingcould instrumentdesignsandallowscientiststoassesstheutilityofthe thushelpachievingthetargetlevelforbiomassmeasurement dataproducts.Thesepackagesmaytrulyberecycledsincethey’re accuracy with a reduced amount of lidar measurements carried to and from the ISS on cargo spacecraft and placed in comparedtowhatwouldberequiredwithoutimager. exterior slots specifically designed for Earth remote sensing ex- periments. Further experiments could be conducted during the However, while sustainability issues are not considered as remaininglifetimeoftheISS(currentexpectation,w10years). crucial,themodel“moredataisbetter”islikelytolastandpriority islikelytobegivenbydefaulttolidarmissionsprovidinghigher 4.2.3. Actionsspecifictothestudycase measurement densities without studying the above mentioned When scientists define mission requirements they tend to issuesand,thus,tothedetrimentofmissionlife-time. favourtheacquisitionofamaximumamountofdata.Foravege- If one of the projects submitted to embed a vegetation lidar tation lidar mission this is liable to reduce the mission life-time, onboardtheInternationalSpaceStationwereapproved,measure- giventhatlaserlife-timeistypicallydrivenbynumberofemitted mentsoftheEarth’sforestssouthof51.6(cid:4)Noverthelifespanofthe pulsesratherthantimeinorbit.Thisiswhy,tooptimizemission ISS would be available. There are several points which argue for specificationswhileadheringtotheprinciples andrecommenda- suchasolution.First,theoperationofanISSlidarwouldobviously tions that mitigate environmental impacts of the mission, we be physically and operationally supported by existing infrastruc- suggestthefollowing: ture. Second, the means to transport the lidar to the ISS and to servicethelidarwhileoperatingaboardtheISSwoulddependona - Identify the best trade-off between measurement contiguity transportationsystemthatisalreadyusedtomaintainthecrewand andlaserlife-timeconsiderations.Thechoicecanthenbemade station. Third, as the ISS was originallyconceived to promote in- whethertogiveprioritytoalongerlifetimeandnoncontiguous ternationalcooperation,weexpectthatthespaceagencieswould measurementsortocontiguousmeasurementswitharesult- bewillingtoworktogetheronanISSlidarsystem,thereforesharing ingshorterlifetime.Duetospaceactivitysustainabilityissues, experiences and costs and avoiding the development of multiple theaddedvalueofcontiguouslidarmeasurementsshouldbe satellitesystemsprovidingduplicateddata.Lastly,whilethesystem carefully examined. Currently, studies that involve regional wouldbedesignedtoacquirerangingandpossiblyactivespectral biomassorcarbonassessmentstendtotreatsatelliteorbitsas measurements on tropical and temperate forests, it would be a observationalunits,i.e.clusters,ratherthanconsideringindi- technologydemonstratorfromwhichlessonscouldbelearnedto vidual laser shots as separate observations (e.g. [59,62,63]. preparefuturespacemissions. Such an approach precludes any need to account for along- track, within-orbit spatial autocorrelation. If laser lifetime is not a concern, then contiguous, along-track measurements 4.3. Thecontributionofairbornemissionstoamoresustainable makesensesincespatialadjacencymakesiteasiertofindand Earthobservationnetwork trackgroundwhileincreasingthenumberofindividualpulse observationsavailabletocharacterizeaparticularcovertypeor 4.3.1. Comparingspaceborneandairbornemissionstooptimize stratum.Iflaserlifetimeisalimitingfactor,thenfewershots theirrespectivecontributions spacedfurtherapartwouldsuffice. The rigid technical specifications associated with space tech- - Explorethelidar signaldynamicinforeststomakesure that nology and related industries might give rise to environmental thepulsecharacteristicsarecapableofsupplyingasignalfrom impacts during system manufacturing (for both the satellite and boththegroundandtheforestcanopyunderawidevarietyof thelauncher),deployment,andoperationalphasesthatwouldbe canopy conditions, from sparse to very dense. Analyses of greaterthanthoseassociatedwithdeploymentofanaircraftsys- ICESat/GLAS data, a large-footprint waveform system, have temabletoprovidedatathatmeetthesamerequirements.Fora highlighted some limitations in forest height assessment in givenmission,then,itwouldbeworthcheckingtoseeifaspace dense forests where an underestimation of height has been systemisthebestsolutionfromanenvironmentalviewpointtaking observed(seee.g.[78]).Sparseforestschallengeanalystspro- into account, among other factors, data requirements, multipur- cessingGLASdataforadifferentreason;thecanopymaybeso poseuses ofdata,andnumberofsatellites launched atthe same diffuse, e.g. high boreal forests, that no discernable canopy time by a given launcher. The results of a comprehensive, signal is returned [67]. Shrublands present a third problem, comparativeanalysisarehardlypredictable.Toourknowledge,no that being the fact that the signal from short-statured paper has been published that indicate trends or that suggest Pleasecitethisarticleinpressas:DurrieuS,NelsonRF,EarthobservationfromspaceeTheissueofenvironmentalsustainability,SpacePolicy (2013),http://dx.doi.org/10.1016/j.spacepol.2013.07.003