How do I set the criteria for anomaly selection? - Strategic ...

The greater the difference of the EMI response of the munitions from that of the ... Detection data for a 105-mm projectile are illustrated below. PageContent SiteNavigation SearchKeyword: Advancedsearch ViewAllSocialMedia Home AboutSERDPandESTCP AboutSERDP AboutESTCP Partnerships InvestmentStrategy SERDP&ESTCPOffice ProgramAreas InstallationEnergyandWater EnvironmentalRestoration MunitionsResponse ResourceConservationandResiliency WeaponsSystemsandPlatforms NewsandEvents Calendar Blog Conferences&Workshops IntheSpotlight FeaturedInitiatives PerandPolyfluoroalkylSubstancesPFASs RangeSustainment EnergyAssuranceandResilience CleanupInitiatives MunitionsResponseInitiatives GreenManufacturingandMaintenance Conservation ToolsandTraining WebinarSeries InstallationEnergyandWater EnvironmentalRestoration MunitionsResponse ResourceConservationandResiliency WeaponsSystemsandPlatforms FundingOpportunities SERDPSolicitations ESTCPSolicitations OtherFundingOpportunities InvestigatorResources SERDPResources ESTCPResources FeaturedInitiatives Per-andPolyfluoroalkylSubstances(PFASs) RangeSustainment EnergyAssuranceandResilience CleanupInitiatives MunitionsResponseInitiatives ClassificationAppliedtoMunitionsResponse GeophysicalClassificationFAQs GeophysicalSystemVerification MunitionsintheUnderwaterEnvironment GreenManufacturingandMaintenance Conservation Home >FeaturedInitiatives >MunitionsResponseInitiatives >ClassificationAppliedtoM... >GeophysicalClassification... GeophysicalClassificationforMunitionsResponseFrequentlyAskedQuestions(FAQs) GeophysicalClassificationAppliedtoMunitionsResponse 1.Whatisdifferentaboutadvancedgeophysicalclassificationcomparedtothingsthathavebeentermedclassificationordiscriminationinthepast? 2.Whatisatargetofinterest? 3.Whattypesofmunitionscanbeclassified? 4.Isgeophysicalclassificationeffectiveifthereareunexpectedmunitionsonthesite? 5.Howdeepcantheadvancedsensorsdetect?Classify? 6.Aretherelimitationsintermsofterrainandvegetationcoverfortheadvancedsensors? 7.Wheredidgeophysicalclassificationnotworkandwhy? 8.Willregulatorsaccepttheuseofthistechnology? 9.Whatresourcesareavailablefordeterminingifadvancedgeophysicalclassificationisappropriatetomysite? 10.HowdoIknowthatacontractorisqualifiedtoconductageophysicalclassificationproject? 11.WhatdoIneedtoknowaboutmysitetodecidewhethergeophysicalclassificationisappropriate? 12.Whatconstitutesaneasy,typical,andhardsite?Whatistypicalclassificationperformanceoneach?Whatistherange? 13.WhendoIusegeophysicalclassificationintheCERCLAprocess? SiteSpecificPlanning 14.DoIneedtodoasite-specifictreatabilitystudy?WhatcantheconclusionsoftheESTCPprogramtellme? 15.DoIneedtodoasurfaceclearancepriortogeophysicalclassification? 16.Ihave20-mmprojectilesonmysite.Willgeophysicalclassificationwork? 17.IfI’mconcernedaboutfuzesandothermunitionscomponents,canIstillusegeophysicalclassification? 18.Willgeophysicalclassificationidentifyallthemunitionsforremoval? 19.Canclassificationtellmethetypeofmunition? 20.Whatdensityofanomaliesistoodensetodoclassification? 21.Whateffectdoesgeologyhaveonclassificationability?HowdoIdeterminewhetherthegeologyonmysiteistoochallengingforclassificationto... 22.Howlongdoesaclassificationprojecttake?Isitsignificantlydifferentthantraditionalprojects? 23.HowdoIestimatepotentialcostsavings? DataCollection 24.Whatisthedifferencebetweenadetectionandcuedsurvey? 25.HowdoIsetthecriteriaforanomalyselection? 26.Whatis“InformedSourceSelection”andhowdoesitwork? 27.Whataretheadvantagesanddisadvantagesofusinganadvancedsensorforthedetectionsurvey? 28.IfIdoadetectionsurveywithanadvancedsensor,canIplantodoclassificationusingthatdata? 29.IhaveanoldmagnetometerorEM61survey.CanIusethatforthedetectionstep?CanIusethatforclassification? 30.WhatcanIexpectforproductionrate? DataAnalysis 31.Whatcausesvariationintheperformanceofanalysts? 32.Howdoestheperformanceofproductioncontractorscomparetodevelopers? 33.Whatwentwrongwiththeanalyststhatperformedpoorly? 34.WhatistheTOIlibrary? QC/QA 35.DoIneedtoseedthesite?WhatseedsshouldIuse?Howmany?Howdeep?WhataccuracyofemplacementanddocumentationdoIneed? 36.Whatproductsanddocumentationcantheprojectteamexpecttoseeduringthecourseoftheproject? 37.HowmuchQCwillberequired? 38.HowmucheffortisrequiredforQAvs.QC? 39.WhocanhelpmewithQAandoversight? 40.Duringthedigging,whatinformationshouldIbecollectingandwhatshoulditbeusedfor? 41.DoIneedtohaveaqualifiedgeophysicistonthegovernmentQAteam? ProjectConclusion 42.Ifwedon’tdigalloftheanomalies,howcanwedemonstratethatclassificationwassuccessfulatmysite? 43.Whataremyoptionsforverificationandvalidation?Howdoestheteamdecidethatthesiteresponseactionisdone? 44.WhattypesofmetaldebrisshouldIexpecttofindonmysiteafteramunitionsresponseusingclassification? 45.Attheendofaproject,canyoutellmewhatthemetalitemsarethatwereleftonthesiteandwhere? AftertheESTCPDemonstrationProgram 46.HowdoIdeterminethecapabilityofnewsensorsthathavenotbeendemonstratedbyESTCP? 47.HowdoIdeterminetheveracityofalternativeclassificationmethods? 48.Whatremainsuntested/unknownattheconclusionoftheprogram? GeophysicalClassificationAppliedtoMunitionsResponse 1. Whatisdifferentaboutadvancedgeophysicalclassificationcomparedtothingsthathavebeentermedclassificationordiscriminationinthepast? Advancedgeophysicalclassificationdescribestheuseofintrinsicpropertiesofaburiedmetalobjecttoclassifythatobjectaseitheratargetofinterestornot.Mostcommonly,theintrinsicpropertiesusedaretheprincipal-axispolarizabilitydecaycurves(sometimescalledEMIfingerprints)whichreflectthesize,symmetry,materialcomposition,wallthickness,etc.oftheobject.Thesepolarizabilitydecaycurvesarerecoveredfromthemeasureddatathroughamodel-matchingprocesscalledgeophysicalinversion.Previousdiscriminationeffortshavebeenbasedonpropertiesofthegeophysicaldatasuchasspatialfootprintoftheanomaly,temporaldecayofanindividualsounding,etc.Whilethesedata-basedparametershavesomerelationtotheburiedobject,theyareconvolvedwithobject’sdepthandorientationrelativetothesensormakingunambiguousanalysisdifficult,ifnotimpossible.Becausegeophysicalclassificationutilizesmoreparametersformakingdeterminationsthatrevealspecific,intrinsicinformationabouttheburieditems,itisinherentlymoreselectivethanpreviouslyutilizedclassificationtechniques.ThisincreaseinselectivityallowsforgeophysicalclassificationtobemoreeffectiveatdiscerningTOIfromnon-TOI. BacktoTop 2. Whatisatargetofinterest? Targetofinterest,orTOI,isthetermusedforitemsthatmustbecorrectlyclassifiedandexcavatedtoaccomplishsiteremediationgoals.Allmunitions,QCandvalidationseeds,andotheritemsdesignatedbythesiteteamsuchassignificantpiecesofmunitions,etc.aretargetsofinterest.SomesiteteamsmayevenwanttoincludecertainfuzesandothermunitionscomponentstotheTOIlist.Munitionsdonothavetocontainhigh-explosivefillertobeTOI;anythingthathastobeexcavatedandexaminedtodetermineifitishazardousis,bydefinition,aTOI. BacktoTop 3. Whattypesofmunitionscanbeclassified? Alltypesofmunitionscanbeclassifiedbuttheefficiencyofcorrectlyclassifyingclutterdependsonmunitionstypesfoundatasite.ThegreaterthedifferenceoftheEMIresponseofthemunitionsfromthatofthepredominantsiteclutter,thegreatertheclassificationefficiency.Largeprojectileslookalmostnothingliketheclutterfoundonmostsitesandcanbeclassifiedwithgreatefficiency.ManycommonclutteritemshaveanEMIresponsesimilartothatof20-mmprojectilessoclassificationislessefficientifthetargetsofinterestinclude20-mmprojectilesandothersmallmunitions. BacktoTop 4. Isgeophysicalclassificationeffectiveifthereareunexpectedmunitionsonthesite? DuringtheESTCPdemonstrations,geophysicalclassificationhasbeenproventobeeffectiveevenwhentherewereunexpectedmunitionsonsite.Intheearlydaysofthedemonstrationprogram,wetestedthisquestionbyseedingthesitewithmunitionsthatwerenotamongthoseknowntobepresentfromthehistoricaldocuments.Alldemonstratorswereabletoidentifytheseunexpectedmunitionsinallcases.Therearetwoaspectstothisquestion.Thefirstcaseinvolvesmunitionsthatarenotexpectedatthesitebutarecontainedinthemunitionslibraryavailabletotheanalyst.Thestandardoperatingprocedurefordataanalysismustinvolveastepwhereresponsesobservedonthesitearecomparedtothefulllibraryofmunitions.Thisprocedurewillidentifyanyunexpectedbutknownmunitions.Thesecond,moredifficultcase,concernsunexpectedmunitionswhoseresponseisnotinthemasterlibraryofsignatures.IntheearlydaysoftheESTCPprogramasthelibrarywasbeingassembled,thiswasnotanuncommonoccurrence.Eventhoughthesignaturesinquestionwerenotinthelibrary,theystillexhibitedallthehallmarksofamunition(size,symmetry,andwallthickness)andthuswereclassifiedasTOI.TheFt.Silldemonstrationsitecontaineditemsdescribedbytheintrusivecrewas“40mmFragBallCupsContainingHE”thatwererelativelythin-walledandthusincorrectlyclassifiedbymostanalysts.Thesitemanagersknewtoexpecttheseitemsbutthisinformationwasneverrelayedtotheanalystsprovingagainthatcommunicationisonekeytoasuccessfulproject.Inbothofthesecases,informationgainedduringthesitesurfaceclearancecanhelpminimizetheproblem.Thesurfaceclearancereportshouldincludeevidenceofanyunexpectedmunitionsencounteredduringtheclearance.ESTCPhasinitiatedaprojecttocompileamasterlibraryofmunitionssignatureswithcompletemetadata.Anumberofordnancemuseumshavebeenvisitedduringthisprojectinanattempttomakethelibraryascompleteaspossibleandminimizethechancesofencounteringunknownmunitionsonaproductionsite.AftercompletionoftheESTCPproject,thelibrarywillbemaintainedandupdatedasnecessarybypersonnelfromtheUSArmyCorpsofEngineersEnvironmentalandMunitionsCenterofExpertise(EMCX).Atthebeginningofamunitionsresponseproject,theGovernmentprojectmanagerwillobtainthecurrentversionofthelibraryanddistributeittotheanalysiscontractor.Anyitemuniquetothesitethatisnotincludedinthemasterlibrarycanbeaddedtothe“site-specificlibrary”tobeusedonthatsite. BacktoTop 5. Howdeepcantheadvancedsensorsdetect?Classify? Forsmallandmediummunitions,TEMTADSandMetalMappercandetecttodepthssimilartothefamiliarEM61.Thetransmitmoment(ameasureoftheexcitationstrengthofthesensor)issomewhatsmallerfortheadvancedsensorsbutthatiscompensatedbytheadditionalreceivechannelsavailablefornoisesuppression.Forlargermunitionssuchas105-mmand155-mmprojectilesandbombs,thedetectiondepthoftheadvancedsensorsisshallowerthananEM61butstilldeeperthanmostitemsarefound.TheadvancedsensorscollectuptotwoordersofmagnitudemoredatathananEM61.Theseadditionaldataalsoallowburieditemstobeadequatelyclassifieddowntotheirdetectiondepth.Thefigureontherightbelowshowstheresponseinchannel2ofanEM61-MK2toa60mmmortarasafunctionofdepth.Forareasonablylow-noisesite,themaximumdepthofreliabledetectionisreachedforamortarinitsleastfavorableorientationatabout65cm.TheplotontheleftshowsthecorrespondingamplituderesponseforaTEMTADS2x2sensor.Here,theamplitude-baseddetectionlimitisreachedatabout58cm.Detectiondatafora105-mmprojectileareillustratedbelow.Thedetectiondepthofa105-mmwithaTEMTADSis~90cmwhileitis~120cmwithanEM61-MK2.Ofcourse,ifthemunitionsofinterestatasiteareexclusivelylarge,deepmunitionsthenamagnetometerremainsthedetectionsensorofchoice.Asmentionedabove,theadvancedsensorscollectmuchmoredatathanthesimplez-axisamplitudeplottedinthetwofigures.TheseextradatacanbeusedinmethodssuchasInformedSourceSelectiontoincreasethedetectiondepthoftheseadvancedsensors. BacktoTop 6. Aretherelimitationsintermsofterrainandvegetationcoverfortheadvancedsensors? Anywherethatthesensorscanbemaneuveredanddatacollectedwithin20to30cmoftheground,classificationcanbesuccessful.OneoftheadvancedsensorscanbeusedanywherethattraditionalsensorsliketheEM61canbedeployed.Showninthephotosbelowarethetractor-borneMetalMappersensor(leftpanel),theTEMTADS2x2inlitter-carrymode(centerpanel),andthehandheldMPVsensorinthewoods(rightpanel). BacktoTop 7. Wheredidgeophysicalclassificationnotworkandwhy? Geophysicalclassificationworksalmosteverywhere.Theprimaryexceptionisthecasewheretheanomalydensityissohighthatthereisnotenoughinformationinthesensorsignalstoresolvetheoverlappingresponsesoftheclosely-spacedobjects.Thiscanoccuratthecenterofheavilyusedtargets,indisposalpitsanddumps,andnearOB/ODareas.Highandvariable(onashortlengthscale)geologicbackgroundmayalsomakeclassificationimpractical,particularlyifthemunitionsofconcernaresmall.IfthebackgroundsignalisasignificantpercentageoftheresponseoftheTOIandvariesonalengthscaleoftensofcentimeters,itwillnotbepossibletoadequatelysubtractthegeologicbackgroundresponseandthepolarizabilitiesderivedwillbeunreliable.Thereareothersitesonwhichitmaynotmakesensetoapplyclassification.OnsiteswithanunusuallyhighratioofTOItoclutter(anESTCPdemonstrationonastrafingrangefellintothiscategory)theremaynotbeenoughclutterdigssavedtomaketheeconomicsofclassificationwork.Similarly,siteswithverysmallnumbersofanomaliesmaynotprovideenoughopportunitiestorepaythecostofmobilizingtheclassificationsensorandcrew BacktoTop 8. Willregulatorsaccepttheuseofthistechnology? Classificationtechnologyhasreceivedregulatoryapprovalforuseonseveralsites.Asmoreandmoreregulatorsgainfamiliaritywithclassification,anincreasingnumberareacceptingitsuseontheirprojects.TheInterstateTechnology&RegulatoryCouncil(ITRC)hascompletedaGeophysicalClassificationforMunitionsResponsetechnicalandregulatoryguidanceandaccompanyingtrainingtopreparestateregulatorstobeinformedparticipantsinsiteteamscontemplatingtheuseofclassification.ESTCPhashadaClassificationAdvisoryGroupsincetheinitiationofthedemonstrationprogram.TheAdvisoryGroup,whichincludesStateandFederalregulators,hasbeeninstrumentalinthedesignandconductoftheESTCPdemonstrations.TheyhavevettedallthemajorconclusionsofthedemonstrationsandbeenheavilyinvolvedintheprogramFinalReport.TheIntergovernmentalDataQualityWorkgrouphaspreparedaGeophysicalClassificationforMunitionsResponse(GCMR)UFP-QAPPtemplateforuseinmunitionsresponseprojects.TheDoDEnvironmentalDataQualityWorkgroup(EDQW)isfinalizingtheDoDAdvancedGeophysicalClassificationAccreditationProgram(DAGCAP)toensurethatperformingcontractorshavetheappropriatequalitysystemsinplaceandtheiremployeeshavethetrainingandexperiencetosuccessfullycompletemunitionsresponseprojectsusinggeophysicalclassification. BacktoTop 9. Whatresourcesareavailablefordeterminingifadvancedgeophysicalclassificationisappropriatetomysite? Thereareseveralsourcesforindependentinformationaboutadvancedgeophysicalclassificationandtheconductofclassificationprojects.TheInterstateTechnology&RegulatoryCouncil(ITRC)GeophysicalClassificationforMunitionsResponseteamhaspublishedTechnologyandRegulatoryGuidanceonthissubjectandisofferingInternet-basedtraining.Thismaterialcanbeaccessedfromthe ITRCwebsite. BacktoTop 10. HowdoIknowthatacontractorisqualifiedtoconductageophysicalclassificationproject? WehavedemonstratedduringtheESTCPprogramthattrainingandexperiencearethebestpredictorsofperformancesuccessonaclassificationproject.Theprimarycriterionforevaluationofcontractorsshouldbesuccessfulpastperformanceonpreviousclassificationprojects.Inaddition,theDoDisintheprocessofestablishingtheDoDAdvancedGeophysicalClassificationAccreditationProgram(DAGCAP).Whenfullyimplemented,onlycontractorsthathavebeenaccreditedbythird-partyAccreditingBodieswillbeeligibletoperformclassificationonmunitionsresponseprojects.Thiswillensurethatperformingcontractorshavetheappropriatequalitysystemsinplaceandtheiremployeeshavethetrainingandexperiencetosuccessfullycompletemunitionsresponseprojectsusinggeophysicalclassification. BacktoTop 11. WhatdoIneedtoknowaboutmysitetodecidewhethergeophysicalclassificationisappropriate? Anup-to-dateConceptualSiteModel(CSM)andaclearanddetailedremedialobjectivearerequiredtoassesswhethergeophysicalclassificationisappropriateonasite.TheCSMshouldcontaininformationontheexpectedmunitions,theirexpecteddepths,anyterrainorvegetationissuesthatwouldimpedeclassification,geologicformationsthatmaycauseinterferencewiththegeophysicsmeasurements,andtheexpecteddensityofanomaliesthroughoutthesite.Allofthesefactorshaveabearingontheefficiencyofclassification.Vagueremedialobjectivessuchas“removeallmunitionstothedepthofdetection”makeitverydifficulttodesignadetectionandclassificationprogramformaximumefficiency.Itisfarbettertodefinethedepthofconcernbasedonanticipatedlandusesandlistthemunitionsexpectedtobeencountered.ThiswillallowtheprojectgeophysicisttodetermineinadvanceiftheremedialobjectivescanbemetusingadvancedEMIsensors. BacktoTop 12. Whatconstitutesaneasy,typical,andhardsite?Whatistypicalclassificationperformanceoneach?Whatistherange? Exampleeasy,typical,andhardsitesfromtheESTCPdemonstrationprogramandtheircharacteristicsaredetailedinthetablebelow.  Characteristics ExampleDemonstrationSites Expectation Easy Limitedmunitionstypes,lowanomalydensity,terrainandvegetationallowforhighqualitydatacollection,nogeologicinterference PoleMountain,WY CampGeorgeWest,COSWProvingGround,AR Almostallanalystscorrectlyclassifyallseeds,eliminate~90%oftheclutterfromthefinaldiglist Typical Mixedmunitions,nonesmallerthan37-mm,lowtomoderateanomalydensity,terrainandvegetationallowforhighqualitydatacollection,moderategeology SpencerRange,TNCampBeale,CA Mostanalystscorrectlyclassifyalltheseeds,eliminate~80%oftheclutterfromthefinaldiglist Hard Oneormoreofmixedmunitions,smallerthan37-mmTOI,highanomalydensity,conditionsmakedatacollectionchallenging,complexgeology MassachusettsMilitaryReservation,MA Ft.Bliss,TXWaikoloaManeuverArea,HI OnlythemostskilledanalystsdetectallTOI,eliminate~50-70%ofclutterfromthefinaldiglist BacktoTop 13. WhendoIusegeophysicalclassificationintheCERCLAprocess? GeophysicalclassificationcanbeusedinanystepintheCERCLAprocesswhereinformationaboutburiedmetalisrequired.ItwillbemostcommonlyusedintheInvestigation(RemedialInvestigationandFeasibilityStudy)andRemediation(RemedialDesignandRemedialAction)phasesoftheCERCLAprocess.Mostofthedemonstrationefforthasbeenfocusedonremedialactionsasthisiswherethebiggestpotentialforsavingsisfound.ESTCPhassupportedseveraldemonstrationsthatemployedclassificationeitherasacomponentoftheRemedialInvestigation(RI)oraspartofaFeasibilityStudyaftercompletionoftheRI.Atsomesites,classificationmaybearequiredcomponentofaRemedialDesigneffort. BacktoTop SiteSpecificPlanning 14. DoIneedtodoasite-specifictreatabilitystudy?WhatcantheconclusionsoftheESTCPprogramtellme? Inmanycasesasite-specifictreatabilitystudyisnotrequired.TheESTCPdemonstrationprogramcomprised26demonstrationsonawidevarietyofsites.Thesimilarityofcluttersizeanddepthandmunitionsdepthsamongthesiteswasstriking.Performancedatafromoneofthedemonstrationswilllikelybeapplicableforpurposesofplanningandcostestimating.Ifthesitepresentsunusualconditions(terrain,mixofmunitions,clutter,geology,etc.)thenatreatabilitystudymaybenecessaryformostaccurateestimationofcostsandperformance. BacktoTop 15. DoIneedtodoasurfaceclearancepriortogeophysicalclassification? Yes,itisrecommended.Thedecisiontoconductasurfaceclearancepriortoclassificationisnodifferentthanthedecisiontoconductasurfaceclearancepriortoanymunitionsresponse.Onmostsites,safetyconcernswilldictateasurfaceclearancebeforegeophysicalinvestigationbegins.Inaddition,itisalmostalwayslessexpensivetopickupthesourceofpotentialclutteranomaliesbeforethegeophysicalworkbeginsthantoprocessthemafterwards. BacktoTop 16. Ihave20-mmprojectilesonmysite.Willgeophysicalclassificationwork? Thelimiteddatainhandshowthatgeophysicalclassificationcanbesuccessfulatsiteswith20-mmprojectilesalthoughtheefficiencyislikelytobelower.ESTCPhasconducteddemonstrationsattwositeswith20-mmprojectilesbutbothhavebeenlessthanidealsites.Asitewith20-mmprojectilesisplannedforthefinalyearoftheprogram.Therearetwothingstokeepinmindwhenconsideringclassificationonasitewith20-mmprojectiles.Thefirstisthattheanomalyselectionthresholdwillhavetobequitelowtoreliablydetect20-mmprojectilestoanyappreciabledepth.Thiswillresultinalargenumberofanomaliesasalotofsmallclutterisalsodetected.Thisisnodifferentthansearchingfor20-mmprojectileswithatraditionalsensor.Thesecondissueisthatclassificationwillnotbeasefficientifthetargetsofinterestinclude20-mmprojectiles.Therearemanymorefragmentsfromlargermunitionsthatlook(inanEMIsense)like20-mmprojectilesthan37-mmprojectiles.Initialindicationsarethatnomorethan60or65%oftheclutterwillbeabletobecorrectlyclassifiedonasitewith20-mmprojectiles. BacktoTop 17. IfI’mconcernedaboutfuzesandothermunitionscomponents,canIstillusegeophysicalclassification? Geophysicalclassificationcanstillbeusedifthetargetsofinterestincludefuzesandothermunitionscomponentsbutatacostinclassificationefficiency.Justasinacasewhere20-mmprojectilesareincludedinthetargetsofinterest,reliablydetectingfuzesandsmallmunitionscomponentsrequiresalowerdetectionthresholdresultinginalargernumberofanomaliestoconsider.ClassificationisalsolessefficientifthesesmalleritemsareconsideredTOIbecausetherearemanymorefragmentsthathaveanEMIresponseclosetothatoffuzesandothersmallmunitionscomponents.Sitemanagersshouldnotplanon80%ormorecorrectclassificationofclutterinthesecases;50to60%ismorelikely.Theprecedingdiscussiondoesnotapplyinthecaseoflargemunitionscomponentssuchasbaseplatesonanartilleryrange.TheselargecomponentswillnotresultindecreasedclassificationefficiencyastheirresponsetoanEMIsensorisfardifferentthanthatofsmallmunitions. BacktoTop 18. Willgeophysicalclassificationidentifyallthemunitionsforremoval? Yes,withinthedetectionrangeoftheinstrument.ThefirstrequirementforsuccessinanESTCPdemonstrationoraproductionuseofgeophysicalclassificationis100%correctclassificationofTOI.TOIincludeallmunitions,anysurrogatesusedasseeds,andanylargemunitionspartsthatthesiteteamdecidesneedtoberemoved.Wehaverepeatedlydemonstratedthatproductiongeophysicistsarecapableof100%correctclassificationofmunitions. BacktoTop 19. Canclassificationtellmethetypeofmunition? Mostclassificationschemesinusetodayarebasedonlibrarymatching.ThederivedEMIresponsesoftheunknownobjectarecomparedtoalibraryofpolarizabilitiesforknownmunitionsitemsand,ifthereisasufficientmatch,anidentificationismade.Themagnitudeofthepolarizabilitiesisafunctionofthesizeofthemunitionsoitwouldbeveryunlikelytoconfusea37-mmprojectilewitha105-mmprojectileforexample.Thereisalwayssomenoiseinfieldmeasurementssoanexactmatchcannotbeguaranteed;itispossibletomistakean81-mmmortarfora75-mmprojectile.Withinsizebands,itisperfectlyreasonabletoexpectidentificationofamunitionusingclassificationmethods. BacktoTop 20. Whatdensityofanomaliesistoodensetodoclassification? Thereisnooneanswertothisquestion;itdependsonanumberoffactors.Onelimitingfactorinanomalydensityistheinformationcontentinthesignalsmeasuredbytheadvancedsensors.Duetoafinitenumberofreceivers,bandwidth,andgeometriccoverage,presentdaysensorsareonlyabletoresolveuptothreetofourindividualanomaliesinthefootprintofthesensor(~1msquare).Thesizeofthemunitionsofconcern,andthusthesizeoftheanomalyfootprint,isalsoafactorintheanomalydensitythatcanberesolved.Classificationhasbeensuccessfulatdensitiesashighas3,000to4,000anomaliesperacrewhenthemajorityofanomaliesaresmallbutusersshouldnotexpecttobeabletoresolveanomalydensitiesmuchhigherthanthat. BacktoTop 21. Whateffectdoesgeologyhaveonclassificationability?HowdoIdeterminewhetherthegeologyonmysiteistoochallengingforclassificationtobereliable?Howdoesthiseffectfrequencyofbackgroundmeasurements? Inallbutthemostextremecases,geologicresponsebyitselfdoesnotimpactclassification.Short-scale(10’sto100’sofcentimetersscale)variationofthegeologicresponseistheproblem.ToproperlyderivetheEMIresponseofaburiedmetalitem,thepartofthemeasuredsignalduetothesensoritselfandthenearbysoilandrockmustbesubtracted.Thisisusuallyaccomplishedbycollectingameasurementatabackgroundlocationthathasnoburiedmetalneartheitemtobeclassified.Thisworkswellaslongasthegeologicresponseatthebackgroundlocationisclosetothatatthelocationoftheunknown.Ifthegeologicresponseislargeandvaryingthenthisbackgroundsubtractionschemenolongerworks.IftheanomalydetectionsurveyisperformedusinganadvancedEMIsensorthemagnitudeandvariationofthebackgroundresponsewillbeapparentinthedynamicdata.Ifaconventionalinstrumentwasusedtoconductthedetectionsurvey,aspotcheckwithtransectscollectedusinganadvancedsensormaybenecessary.Ifthereisanyquestionifthegeologyisasignificantfactorataparticularsite,asmalltreatabilitystudyisprobablyneeded.Moderategeologicresponsemaynotimpactthefrequencyofbackgroundmeasurementsbutwillnecessitateincreasedattentiontothespatialdistributionofbackgroundlocations.Thefieldgeophysicistsshouldstrivetocollectabackgroundmeasurementincloseproximitytoeachunknownmeasurementtoguardagainstgeologicvariation. BacktoTop 22. Howlongdoesaclassificationprojecttake?Isitsignificantlydifferentthantraditionalprojects? Aclassificationprojectcanoftenbesignificantlyshorterthanthesameprojectconductedusingtraditionalmethods.Theinitialdetectionsurveymaytakelongerusingadvancedsensors,particularlyifthesitelendsitselftouseofarraysoftraditionalsensors.Afterthat,everyaspectofaclassificationprojectisfaster.Ateamcollectingcueddatacancueroughlytwicethenumberofanomaliesadaythatanintrusivecrewcanprosecute.Onmanysites,80%ofthedigscanbeavoidedusingclassification.Overthecourseofalargeprojectthesefactorsadduptoalargetimesavingusingclassification. BacktoTop 23. HowdoIestimatepotentialcostsavings? Onatypicalsitewhere80%ofthecluttercanbecorrectlyclassified,costsavingsof40%to50%arepossibleusingclassification.ImplementationofclassificationdoesrequireextracostsforadditionaldatacollectionandanalysisandtheQCrequiredforaqualityprojectbutthesavingsaretheresultofclutterdigsavoided.Agoodruleofthumbis$125to$200savedperavoideddig.Siteswitheasysoilandshallowclutterwillhavealowercostperdigandthosewithtougherconditionsanddeeperclutterwillhaveahighercost.IntheESTCPdemonstrations,detectionsurveycostshavebeenaboutthesameasasurveyusinganindividualEM61,cueddatacollectionandanalysishasbeenabout$35peranomaly,andtheadditionalQCrequiredbythisapproachhasbeenabout10%ofprojectcosts. BacktoTop DataCollection 24. Whatisthedifferencebetweenadetectionandcuedsurvey? Adetectionsurvey,sometimesreferredtoasadynamicsurvey,isa100%-coveragegeophysicalsurveyofasite.Adetectionsurveyisoftenaccomplishedusingbackandforthparallelsurveytracks(akintomowingthegrass)withthetrackspacingafunctionofthesensorwidth,smallestmunitiontobedetected,andanomalydensity.Thedatafromthegeophysicalsensorarecombinedwithgeolocationdata(usuallyGPSdatabutothergeolocationsystemsareusedundertreecover)andmapped.Detectionsaredeclaredatthelocationsofanomalousgeophysicalresponsecomparedtobackground.AdetectionsurveycanbeconductedusingeitheratraditionalEMIsensororoneoftheadvancedEMIsensorsdevelopedforclassification.UseofanadvancedsensorforthedetectionsurveyallowsformoresophisticateddataanalysessuchasInformedSourceSelection.Acuedsurveyisastationarydatacollectionoverapreviously-identifiedanomalyforthepurposeofacquiringhigher-qualitydatatouseforclassification.CuedsurveysonlymakesenseusinganadvancedEMIsensor;notenoughinformationisavailableinasingle-coil,single-axissensorsuchastheEM61-MK2tomakecueddatacollectionworthwhile. BacktoTop 25. HowdoIsetthecriteriaforanomalyselection? AnomalyselectioncriteriashouldbesetbasedonthesiteremedialactionobjectivesandCSM.Oncethemunitionsofconcern(TOI)andthedepthofconcernarespecified,itisstraightforwardtodeterminewhichcombinationofmunitionanddepthresultsinthelowestamplitudeanomalyandsetthatastheamplitudethresholdforanomalyselection.Ifresponseamplitudeisthesolecriterionbeingusedthentheprocessisalmostcomplete;theonlyremainingtaskistocomparethisamplitudethresholdtomeasuredsitenoise.IfthisminimumamplitudeisgreaterthanfivetimestheRMSnoiseatthesitethentheremedialobjectivesareachievablewiththesensorselectedandgeophysicalsurveyscanproceed.If,however,theminimumamplitudeiswithinthesitenoisetheanomalyselectionamplitudewillhavetobeincreasedandthesiteteamwillhavetodecideiftheycantoleratethereduceddepthofdetectionthatresults.Ifadditionalanomalyselectioncriteriainadditiontoamplitudewillbeemployedthentheamplitudethresholdisonlythefirststep.IfanomalydetectionisbeingperformedusingatraditionalsensorsuchastheEM61-MK2,thechoicesforadditionalcriteriaarelimited.AnalystshavehadsomesuccessinthepastusingcriteriasuchasarealextentoftheanomalousresponseandsignaldecayofthepeaksoundingtoremoveanomaliesthatcouldnotpossiblybeduetoaTOIfromfurtherconsideration.IftheanomalydetectionsurveywasconductedwithanadvancedEMIsensorthenthepossibilitiesforadditionalanomalyselectioncriteriaaremuchlarger.Theseadditionalcriteriacanrangefromuseofadditionalreceivechannelstodiscriminateagainstnoisespikesatthesimplesttoafullanalysisofalldatacollectedovereachanomalytorejectsmall,shallowclutterthatcouldnotpossiblybeTOI.ThislatterprocessistermedInformedSourceSelection. BacktoTop 26. Whatis“InformedSourceSelection”andhowdoesitwork? “InformedSourcedSelection”referstotheuseoftheextrainformationinherentinthesignalsfromadvancedEMIsensorstoselectinthedetectionsteponlythoseburiedmetalsourcesthatcouldbecausedbyatargetofinterestforfurtherconsideration.Notethatthistechniquefocusesontheburiedmetalitems(sources)thatresultintheanomalyratherthantheanomalyitself.TheremedialobjectiveistoremoveTOI(thatis,sources)ratherthananomalies.Traditionalanomalyselectionhasbeenbasedontheamplitudeoftheresponseobservedduringageophysicalsurvey.Iftheremedialobjectiveistoidentify37-mmprojectilesatonefoot,theknownminimumEM61signalingate2forthattargetis5.2mV.Thus,allanomalieswithpeaksignalabove5.2mVwouldbeselected.Thismethoddefinitelyidentifiesallanomaliesresultingfrom37-mmprojectilesdowntoonefootbutitalsoidentifiesalargenumberofanomaliesduetomuchsmallerclutternearthesurface.InsomecasessecondaryfilterssuchasarealextentoftheanomalyanddecayrateoverthefourchannelsoftheEM61-MK2areusedtotrimthelistbutthesehaveproventobeonlypartiallysuccessful.TheadvancedEMIsensorscanbeconfiguredtoexcitethetargetalongmultipleaxesandsensetheinducedfieldsalongthreeaxesformuchlongertimes.Thisprovidesatleastanorderofmagnitudemoreinformationtouseforsourceselectioncomparedtoatraditionalsensor.Thisextrainformationcanbeusedtodiscriminateagainstnoisespikescausedbyenvironmentalinterference,lessentheinterferencecausedbysitegeology,anddiscriminateagainstanomaliescausedbysmall,shallowclutter.InformedSourceSelectionhasbeenshowntoreducethenumberofitemsthatrequirecueddatacollectionbyuptoafactorofthreewithnoimpactonthedetectionofmunitions. BacktoTop 27. Whataretheadvantagesanddisadvantagesofusinganadvancedsensorforthedetectionsurvey? Therearetwobigadvantagestousinganadvancedsensorforthedetectionsurvey.Thereceivercubesintheadvancedsensors(8-or10-cmcubes)aremuchsmallerthanthe50-x100-cmloopreceiverinatraditionalsensor.This,coupledwiththeprecisepositionandorientationsensorsontheadvancedsystems,resultsinmuchbetterlocationoftheobservedanomalies–oftenwithin15to20cmasopposedto50to75cmwithtraditionalsensors.Thesecondadvantageistheabilitytoperformmuchmoresophisticatedanomalyselection(e.g.InformedSourceSelection)whenworkingwithdatafromtheadvancedsensors.TheadditionalinformationprovidedbytheadvancedEMIsensorsaffordstheanalysttheopportunitytousemorethantheobservedsignalamplitudetoselectonlythoseanomaliesthatcouldresultfromatargetofinterestforfurtherconsideration.Atpresent,thedisadvantagetotheuseofadvancedsensorsfordetectionsurveysistheproductionrate.Thelimitedsizeandinabilitytoconfigurethemasarrayslimitsthedailysurveycoverageto0.75to1.5acres.Onsmallsitesthisisnotanissuebutonlarger,opensites,itiscommontomakeanarrayofEM61sandsurveyupto5acresperday.Dependingonsiteconditions,thislimitedsurveycoveragecanbeoffsetbythereductioninthenumberofanomaliesthatrequiredcueing.Onasitewithmoderatelyhighanomalydensity(1000anomaliesperacreorabove)mostlyconsistingofsmallclutter,InformedSourceSelectioncanreducethenumberofcuedmeasurementsbyuptoafactorofthree.Onsiteswithlowanomalydensitiessuchasbufferareas,thelimitednumberofanomaliesdoesnotaffordenoughreductionincueddatacollectionefforttooffsetthelowersurveyratesoftheadvancedsensors. BacktoTop 28. IfIdoadetectionsurveywithanadvancedsensor,canIplantodoclassificationusingthatdata? Advancedsensorsblurthelinebetweendetectionandclassificationphasesofamunitionsresponseproject.Surveydatacollectedusinganadvancedsensorcanbeusedinmultiplewaysalongthecontinuumfromdetectionthroughclassification.Totakeadvantageofthis,thesensormustbeconfiguredsothattheburiedmetalobjectisinterrogatedfromenoughdifferentdirectionstoallowextractionofreliableresponsecurves.ThisismoreeasilyaccomplishedwiththeTEMTADS2x2sensoratpresentthantheMetalMapperbutvariantsofbothsensorsarebeingdevelopedthatwillmakethismoreroutine.Therearethreepossibilitiesfortheuseofdynamicdata.Inthefirst,sometimesreferredtoas“InformedSourceSelection,”aninitialanalysisisperformedforeachanomalyandanyanomaliesthatcouldnothavebeencausedbyatargetofinterestatthesiteareremovedfromtheanomalylist.Thisprocesscanradicallyparethenumberofanomaliesthatrequirefurtherinvestigationbyeliminatingthosethatresultfromsmall,near-surfacefragmentsthatgivealargesignalduetotheirshallowdepth.DuringtheSanLuisObispoTreatabilityStudythisprocedurewasusedtodecreasethenumberofanomaliesthatrequiredcueingbyafactorofthree.Detailsofthisareinthetablebelow.StatisticsfromSanLuisObispoDemonstration AmplitudeResponseAnomalies 16,202 InformedSourceSelection–AnomaliesforCueing 7,035 TotalDigs(5.1%ofAmplitudeResponseAnomalies) 828   Can’tAnalyze 104   PotentialTOI(including169QCandvalidationseeds) 493   Training 148   VerificationandValidation 83 Thesecondwaytousedynamicdataistomakeclassificationdecisionsonanomaliesthatareclearlyclutterortargetsofinterestusingthedynamicdataandonlycollectingcueddataoverthosethatareambiguousfromthedynamicdata.Thiscanresultinanevengreaterreductioninthenumberanomaliesrequiringcueddatacollection.Finally,onsiteswithmoderateanomalydensityandlowgeologicbackground,itmaybepossibletoperformclassificationusingthedynamicdataonly.TheexamplebelowisfromtheESTCPdemonstrationatSpencerRange.Thepanelontheleftshowstheresultsachievedanalyzingthecueddata.TheseareveryimpressivewiththeinitialpartofthecurvealmostverticalindicatingveryefficientclassificationofTOIandasharpbreakafter100%oftheTOIhadbeenidentified.Morethan85%oftheclutterwascorrectlyclassifiedinthisanalysis.Thepanelontherightshowstheresultsachievedanalyzingthedynamicdataonly.TheclassificationofTOIisnotasefficientinthisanalysisasindicatedbytheslantoftheinitialpartofthecurve.Additionally,only75%oftheclutteriscorrectlyclassifiedbytheanalystmeaningabout30extradigswouldberequiredatthissmallsite.Thecostoftheseextradigsislessthanthecueddatacollectionstepwouldhaverequired. BacktoTop 29. IhaveanoldmagnetometerorEM61survey.CanIusethatforthedetectionstep?CanIusethatforclassification? ItispossiblethatanexistingmagnetometerorEM61surveycanbeusedforanomalydetectioninaclassificationprojectbutitisnotlikely.TheexistingdatamustbecloselyscrutinizedtoseeiftheymeetprojectDQOsbeforetheirusecanbeconsidered.Evenifthelanespacinganddetectionthresholdwereappropriateforthenewusetheunderlyingpositioningprecisionoftheexistingsurveymayleadtoaninefficientclassificationprocess.Positionuncertaintiesinanomalylocationgreaterthan20to30cm(eithercausedbyinaccuratepositioningsystemorexcessivelanespacing)greatlyincreasethecostofacueddatacollectiononallbuttheleastdensesites.Onasitewithnormaltohighanomalydensity,theextracostofaprecisely-positionedsurveywilllikelyberepaidduringcueddatacollection.MagnetometerandEM61datacannotbeusedforclassification.Atanyreasonabledistancefromthesensor,themagnetometeranomalyobservedcontainsnoinformationaboutshapeanddoesnotlenditselftoclassification.EarlydemonstrationsintheESTCPseriesshowedthatEM61isunsuitableforefficientclassification. BacktoTop 30. WhatcanIexpectforproductionrate? BoththeTEMTADS2x2andMetalMappersensorscanroutinelyachieveratesof150to200cuedinterrogationsperday.TheTEMTADSismoremaneuverablesoachievesthehighestproductivitywhentheanomaliesareclosetogether;theMetalMapperinitstractor-borneconfigurationmovesfastersoachievesthehighestproductivitywhentheanomaliesaremorewidelyspaced.TheadvancedEMIsensorsarenotoptimizedfordeploymentasarrayssothesurveyproductivityislimitedtowhatcanbeachievedwithasinglesensor.Insurveymode,theMetalMapperroutinelycovers1to1½acresperdayandtheTEMTADS2x2covers¾to1acreperday. BacktoTop DataAnalysis 31. Whatcausesvariationintheperformanceofanalysts? OverthecourseoftheESTCPdemonstrations,wehaveobservedthatthereisastrongcorrelationofanalysttrainingandexperiencewithabilitytocorrectlyclassifyclutter.Althoughthereissomevariationwithinthebands,novicestendtoperformworstandthosewithexperienceonmanyprojectsalmostalwaysachievenear-perfectresults.Almostallanalystswereabletocorrectlyclassify100%ofthetargetsofinterest(onlyahandfulofanalyststhatneverseemedtounderstandthemethodmissedTOI)buttheefficiencywithwhichtheywereabletocorrectlyclassifycluttervariedwithexperience.TheplotbelowillustratesthisfortheESTCPSpencerdemonstration.Thebest-performinganalyst,andoneofthemostexperienced,onlyrequired53clutterdigstoidentifyalltheTOIwhileoneofthenovicesrequiredover800.OneimportantcomponentoftheDoDAdvancedGeophysicalClassificationAccreditationProgram(DAGCAP)istherequirementforaccreditedorganizationstoproperlytraintheiremployees,maintainrecordsofthattraining,anddocumentresultsofeachemployeesdemonstrationofcapabilitiesforthetaskstheyareassigned.Thiswillallowsiteteamstobeconfidentthatskilledanalystsaredeployedtotheirprojects. BacktoTop 32. Howdoestheperformanceofproductioncontractorscomparetodevelopers? Atallbuttheeasiestsites,thedevelopersoftheclassificationsoftwareweremoreeffectiveinremovingclutterthantheproductiongeophysicistsbutnotbyenoughtocompromisetheremediationgoals.AnexamplefromtheESTCPdemonstrationattheformerSpencerRange,TNisshownbelow.TheleftpanelshowstheperformanceofoneofthedeveloperswhenanalyzingMetalMapperdata.Afterabout50digsfortraining,theresultsarealmostperfect.Thisanalystidentified100%ofthetargetsofinterestwhilecorrectlyclassifyingmorethan90%oftheclutter.Theproductiongeophysicist(rightpanel)alsoidentified100%ofthetargetsofinterestwhilecorrectlyclassifying80%oftheclutter.Thislevelofperformance(80%clutterreduction)isthebasisforallcostprojectionsinthisdocument. BacktoTop 33. Whatwentwrongwiththeanalyststhatperformedpoorly? Althoughanalysisofclassificationdataisnotinordinatelydifficult,itdoesrequirealevelofcareandattentiontodetailabovethatrequiredfortheroutineanalysisofEM61data.Mostofthehandfulofanalystswithverypoorperformancedidnotpayenoughattentiontotheprojectuntilnearwhentheiranalysiswasdue.Then,iftheywereconfused,therewasnotimetoseekhelportheyrushedthroughtheanalysisatthedeadline.DetailedSOPswithrequiredQCcheckswouldhavetakencareofmostofthepoorperformanceintheESTCPdemonstrations.Theoneortworemainingproblemsresultedfromthewrongemployeeinthewrongplace.ThetraininganddocumentationrequirementsintheDoDAdvancedGeophysicalClassificationAccreditationProgramaredesignedtoeliminatethisproblem. BacktoTop 34. WhatistheTOIlibrary? TheTOIlibraryreferstoacollectionofresponses(polarizabilitydecaycurvesorEMIfingerprints)correspondingtocommonlyoccurringmunitionsitems.MostclassificationschemesinusetodayinvolvecomparingtheEMIresponsesoftheunknownobjectstoeachentryinthelibraryandusingthematch(orlackofmatch)todecideiftheunknownislikelyamunitionorclutter.ESTCPiscompilingamasterTOIlibraryofmunitions’responsesalongwithcompletemetadata.ThislibrarywillbemaintainedandupdatedasrequiredbytheUSArmyCorpsofEngineersandhostedonaGovernmentsite.ThecurrentversionofthelibrarywillbedownloadedbytheGovernmentprogrammanageratthestartoftheprojectanddistributedtothegeophysicalcontractor.Formostprojects,thiswillbetheonlylibraryneeded.Somesitesmayhaveuniquemunitionsspecifictotheformermissionand,inthosecases,themasterlibrarywillneedtobesupplementedwithadditionalsite-specificresponses.Proceduresforconstructingthesesite-specificlibrarieswillbedescribedintheon-linehelpforUX-Analyze. BacktoTop QC/QA 35. DoIneedtoseedthesite?WhatseedsshouldIuse?Howmany?Howdeep?WhataccuracyofemplacementanddocumentationdoIneed? Yes,allmunitionsresponseprojectsshouldincludeblindseeds.How,andtowhatextent,toemplaceseedswillbeasiteteamdecision.ArobustQC/QAprogramisfoundedonblindseeds.QCseedswillbeemplacedbythecontractorasonecomponentoftheirqualityprogramandshouldbeblindtothefieldgeophysicistsandanalystsasspecifiedinthecontractor’sapprovedfirewallplan.Validation(orQA)seedswillbeemplacedbytheGovernmentoritsrepresentativeandshouldbeblindtoallcontractorpersonnel.Thecontractor’sperformanceonboththeQCandvalidationseedswillbeoneofthelinesofevidencethatthesiteteamwillrelyonwhenassessingprojectresults.InmanycasestheseedswillbepredominantlymunitionssurrogatessuchasIndustryStandardObjects(ISOs).ISOs( ESTCPGSVreport)havetheadvantageofbeinginexpensiveandreadilyavailableand,sincetheyhavebeenusedatmanysitesacrosstheUS,itiseasytocomparetheirresponsetothatobservedatothersites.Sometimes,thesiteteamorthepublicwillfeelmorecomfortableifsomeoftheseedsareinertversionsofthemunitionsexpectedonthesite.WhateverischosenasseeditemsshouldnotbesodifferentfromtheTOIexpectedatthesitethattheseedsdrivetheanomalyselectioncriteriaorclassificationthresholds.Correspondingly,theseedsshouldnotbesoeasytoidentifythattheydonoteffectivelytestthecontractor’sprocedures.Oneoftheprimarypurposesofablindseedingprogramistoprovidecontinualconfirmationthatprojectqualityisbeingmaintained.Emplacingenoughseedsthateachteamwillencounteraseedeachdaywillensurethis.Ifthesurveycrewcancoveroneacreperdayindynamicmode,thiswillresultinoneseedperacre.Ifthecuedteamcancue200anomaliesperdaythenoneseedper200anomalieswillbeneeded.Thismaybemorethanoneperacredependingonanomalydensity.Finally,iftheintrusivecrewcanrecover100itemsperday,therewillneedtobeoneseedper100digsexpected.Thepointoftheseedprogramistoconfirmquality,notinvestigatesensorcapabilities.Allseedsshouldbeemplacedatadepthatwhichtheyareexpectedtobedetectedandclassified.Otherwiseitisdifficulttodrawconclusionsfromamissedseed.Thiswillnormallybewithinthedepthofconcernintheremedialobjectives.AtESTCPdemonstrations,theseedshavebeendistributeddowntothedepthofconcernwithmostoftheseedsmatchingthedepthwherethemunitionisexpected.FortheSanLuisObispoTreatabilityStudythismeantthe37-mmsurrogateswereemplaceddownto30cmwiththemajorityoftheseedsinthe4to12cmrangeatwhich37-mmprojectileshadbeenobservedintheRI.SeedlocationsshouldbemeasuredwithaprecisiongreaterthanthespecificationsforseedlocationaccuracyintheQAPP.Usuallythiswillmeancm-levelGPSprecision.ItisgoodpracticeforthecontractortoemplacetheirQCseedssothattheyfullycovertheremedialdepthandconditionsexpectedontheproject;theymayevenbiasthedepthstothelowerendoftheremedialobjective.Thiswillallowthecontractortouncoveranyprocessissuesthatmaycausethemtobeunabletoaccomplishtheremedialobjectives.Thevalidationseedswillnormallybeemplacedatdepthsmoreinlinewiththeexpecteddepthatthesiteastheconsequencesofavalidationseedfailurearesignificant. BacktoTop 36. Whatproductsanddocumentationcantheprojectteamexpecttoseeduringthecourseoftheproject? ThefirstdocumentastakeholdershouldexpectistheprojectQAPP.Amongotherthings,theQAPPwilllisttheDataQualityObjectives,thecorrespondingMeasurementPerformanceCriteriaandMeasurementQualityObjectives,andthepreliminaryValidationandVerificationplans.Astheprojectproceedsasurfaceclearancetechnicalmemorandum,aQCseedreport,andanIVSmemorandumwillbesubmittedalongwithweeklyQCsummaryreports.Attheconclusionofallgeophysicalmeasurements,arankedanomalylistwillbesubmittedwithapreliminarystopdigpoint.Thiswillbeaccompaniedbyasummarymemothatdescribestheclassificationapproachused,thevariousthresholdsemployedintheranking,andanyotherdetailsnecessarytodocumentthedecisionsmade.Attheconclusionofintrusivework,thecontractorwillsubmitthefinalrankedanomalylist,theresultsofallintrusiveinvestigationsandaproposedfinalvalidationandverificationplan. BacktoTop 37. HowmuchQCwillberequired? SiteteamscontemplatingaprojectusingclassificationshouldplanforextraQCandQAeffortcomparedtoatraditionalproject.PartofthisextraexpenseisduetothelevelofconfidencerequiredtoleavesignificantmetalobjectsinthegroundandpartisduetotheadditionalqualitychecksthatarepossibleasthefieldtransitionstotheuseofGCMR-QAPPs.BasedonahandfulofESTCPdemonstrations,wehavefoundthatabout10%ofthetotalprojectfundinghasbeendevotedtoQCwithasmalleramountrequiredforGovernmentQA.AscontractorsgainexperiencewithclassificationandtheQCrequiredandtheanalysissoftwareevolvestobettersupporttheseneedsthiscostwillcomedownbutwilllikelyalwaysbehigherthanprojectsinthepast. BacktoTop 38. HowmucheffortisrequiredforQAvs.QC? QArequireslesseffortthanQCduringmostphasesoftheproject.Aspartofthesitepreptask,thecontractorQCgeophysicistwillprepareasurfaceclearancetechmemo,aseedemplacementreportandanIVSreport.EachofthesewillrequireQAeffort.Duringthedatacollectionphasesoftheproject,theQCgeophysicistwillhavedailyQCcheckstoperformandwillcombinethesedailyreportsintoaweeklyQAsubmission.ThegovernmentorstakeholderQAgeophysicistmaychoosetoexaminethesesubmissionsingreatdetailforthefirstoneortwoweeksbutafterthattheburdenshouldbelower.ThegovernmentQAgeophysicistfortheSanLuisObispodemonstrationreportsspendingabout½dayoneachweeklyQAreport.AstheanalysissoftwarematuresandtheQCandQAreportformatsarestandardized,theeffortrequiredforbothQCandQAwilllikelybereduced. BacktoTop 39. WhocanhelpmewithQAandoversight? USArmyCorpsofEngineersprogrammanagerscancontactUSACEgeophysicistsforhelpwithQAandoversight.ArmyEnvironmentalCommandandNavalFacilitiesEngineeringCommandPMscanexpectsomeassistancefromheadquartersbutmayneedtocontractforthird-partyQAhelp.Somestateshavehiredthird-partyQAconsultantsaswelltohelpthemwithoversightongeophysicalclassificationprojects.Thisisanoptionforstakeholdersassociatedwithamunitionsresponseinvolvinggeophysicalclassification. BacktoTop 40. Duringthedigging,whatinformationshouldIbecollectingandwhatshoulditbeusedfor? Ataminimum,theintrusivecrewshouldrecordthedepthandsizeofallrecovereditemsalongwithagoodphotographandaprecisepositionmeasurement.Thesedatawillbeusedbytheanalysttoconfirmthattherecovereditem(s)areconsistentwiththegeophysicaldata.Thepositionmeasurementcanbeusedtoconfirmthatanincorrectnearbyitemwasnotdug. BacktoTop 41. DoIneedtohaveaqualifiedgeophysicistonthegovernmentQAteam? Yes.ItisawisepolicytohaveaqualifiedgeophysicistontheGovernmentQAteam.Ageophysicistwillbeabletoquicklyverifythecorrectnessofdatacollectionandanalysisreports,judgethecompletenessofroot-causeanalyses,andrenderaninformedjudgementontheadequacyofanyproposedcorrectiveactions.Duringtheverificationandvalidationphase,theQAgeophysicistwillbeabletoquicklychooseitemsforvalidationdigs. BacktoTop ProjectConclusion 42. Ifwedon’tdigalloftheanomalies,howcanwedemonstratethatclassificationwassuccessfulatmysite? Anumberoflinesofevidencewillcontributetoconfidencethatclassificationwassuccessfulataparticularsite.Throughouttheproject,QCchecksdescribedinWorksheet22oftheGCMR-QAPParebeingmadedailyandQAchecksarebeingperformedweekly.PerhapsthemostimportantofthesearethesuccessfuldetectionandcorrectclassificationofboththeQCandvalidation(orQA)seeds.IfalltheMeasurementQualityObjectives(MQOs)havebeensatisfiedthroughouttheprojectthenthesiteteamcanhaveconfidencethattheMeasurementPerformanceCriteriahavebeensatisfiedandconsequentlytheDataQualityObjectiveshavebeenmet.AttheconclusionoffieldworktheverificationandvalidationproceduresoutlinedinWorksheets34and35oftheQAPPwillbeperformedasthefinalevidenceofsuccess.Formostprojectsthesewillincludeaverificationthatthecorrectstop-digthresholdhasbeenchosenandvalidationoftheentireclassificationprocess.Verificationofthestop-digthresholdcanbeaccomplishedbysimplydiggingsomenumberofanomaliespastthecontractor’sproposedstop-digthreshold.ThebetadraftoftheGCMR-QAPPtemplaterecommendsdigging200anomaliespastthelastTOIencountered.Ifthecontractorchoosesaveryconservativethresholdthenno(orfew)extradigsmayberequired.Alternatively,ifthecontractorisaggressiveandonlyrecommendsdiggingahandfulofanomaliespastthelastTOIthen200moredigswillberequired.Validationoftheclassificationprocesscanalsobeaccomplishedrelativelyeasily.ThesuccessfuldetectionandclassificationofallblindseedshasalreadyshownthatgivenaccurateEMIresponses(features)thatcorrectdecisionscanbemadeandthattheintrusivecrewcanrecoverthecorrectitem.Allthatisleftistoshowthatthefeaturesthatwerethebasisforthedecisiononeachitemwerecorrect.Excavationofanumberofanomalies(theGCMR-QAPPtemplaterecommends200foralargeproject)thatwereclassifiedasnot-TOIandcomparisonofthephysicalsizeandshapeoftherecovereditemwiththeanalyst’sbasisforherclassificationdecisionwillprovideconfidencethattheclassificationdecisionswerebasedonanaccuratedescriptionofeachitem. BacktoTop 43. Whataremyoptionsforverificationandvalidation?Howdoestheteamdecidethatthesiteresponseactionisdone? Verificationandvalidationactivitiesattheconclusionofaprojectareonlytwoofthemanylinesofevidencethatwillgivesitemanagersconfidencethatresponseatthesiteis“done.”ThroughouttheprojecttheQCchecksbeingmadedailyandtheweeklyQAchecksarealsoveryimportantconfidence-buildingactivities.TheseincludethesuccessfuldetectionandcorrectclassificationofboththeQCandQA(orvalidation)seeds.IfalltheMeasurementQualityObjectives(MQOs)havebeensatisfiedthroughouttheprojectthentheverificationandvalidationproceduresoutlinedinWorksheets34and35oftheQAPPareallthatremain.Formostprojectsthesewillincludeaverificationthatthecorrectstop-digthresholdhasbeenchosenandvalidationoftheentireclassificationprocess.Verificationofthestop-digthresholdcanbeaccomplishedbysimplydiggingsomenumberofanomaliespastthecontractor’sproposedstop-digthreshold.ThebetadraftoftheGCMR-QAPPtemplaterecommendsdigging200anomaliespastthelastTOIencountered.Ifthecontractorchoosesaveryconservativethresholdthenno(orfew)extradigsmayberequired.Alternatively,ifthecontractorisaggressiveandonlyrecommendsdiggingahandfulofanomaliespastthelastTOIthennearly200moredigswillberequired.Validationoftheclassificationprocesscanalsobeaccomplishedrelativelyeasily.ThesuccessfuldetectionandclassificationofallblindseedshasalreadyshownthatgivenaccurateEMIresponses(features)thatcorrectdecisionscanbemadeandthattheintrusivecrewcanrecoverthecorrectitem.Allthatisleftistoshowthatthefeaturesthatwerethebasisforthedecisiononeachitemwerecorrect.Excavationofanumberofanomalies(theGCMR-QAPPtemplaterecommends200foralargeproject)thatwereclassifiedasnotTOIandcomparisonofthephysicalsizeandshapeoftherecovereditemwiththeanalyst’sbasisforherclassificationdecisionwillprovideconfidencethattheclassificationdecisionswerebasedonanaccuratedescriptionofeachitem. BacktoTop 44. WhattypesofmetaldebrisshouldIexpecttofindonmysiteafteramunitionsresponseusingclassification? Thevastmajorityofmetalitemsonatypicalmunitionssitearethumb-sizedfragmentsandsmaller.ThephotographbelowshowsacollectionoffragmentscollectedduringarecentESTCPdemonstration.Ifthequestionisposedas“Whatistheaverage(ormedianortypical)metaldebrisonmysite?”thentheanswerisinthephotograph.Thereareasmallnumberoflarger,recognizablefragmentsonmostsites.Itemssuchasfinassembliesonmortartargets,baseplatesonartilleryranges,boxfinsonpracticebombtargets,etc.arenormallyclassifiedasnottargetsofinterestandleftonthesite.IncaseswhereparticularsiteuseorstakeholdersensitivityrequirestheremovaloftheselargeritemstheycanbeaddedtothelistofTOI. BacktoTop 45. Attheendofaproject,canyoutellmewhatthemetalitemsarethatwereleftonthesiteandwhere? Yes.Themasterprojectdatabasedeliveredbythecontractorwillhaveanentrycorrespondingtoeveryitemclassifiedasclutterwiththeitem’sEMIsignatureandroughsize.Becausetheclassificationprocessisdesignedtoonlyleaveitemsbehindthatcanbeaffirmativelyclassifiedasclutter,muchmoreisknownaboutthesiteafteraremedialactionusingclassificationthenafteratraditionalproject. BacktoTop AftertheESTCPDemonstrationProgram 46. HowdoIdeterminethecapabilityofnewsensorsthathavenotbeendemonstratedbyESTCP? Thesensorsthathavebeenusedsuccessfullytodateareallmulti-axis,multi-coilelectromagneticinduction(EMI)sensors.OthersensormodalitiesmaybedevelopedtoperformtheclassificationfunctionbutanynewsensorwillhavetoshareseveralcharacteristicswiththeadvancedEMIsensors.Theprimarythresholdforanewsensoris“noblackboxesallowed.”Withoutatransparentprocess,stakeholderswillnotbeabletojudgethevalidityofaclassificationproject.Thismeansthatthepurveyorofthenewsensorshouldbeabletoprovideadetailed,writtendescriptionofthesensingmodalityandanalysisapproach,predictinadvancethesignalsandcharacteristicsexpectedfromthemunitionsitemslikelytobeencounteredonthesite,andhavedemonstratedtheperformanceofthenewsensoronatestsiteorpreviousproject. BacktoTop 47. HowdoIdeterminetheveracityofalternativeclassificationmethods? Theprimarymeasureoftheusefulnessofanyclassificationmethodisdemonstratedperformance.Resultsfromablindtestsite,analysisofthedatafromapriorESTCPdemonstration,orapreviousprojectwillallowaprospectiveusertojudgethereliabilityofaproposedmethod.SuccessfuldetectionandcorrectclassificationofallTOI,includingallseeds,andthefractionofcluttercorrectlyclassifiedarethemeasuresthatshouldbeexamined. BacktoTop 48. Whatremainsuntested/unknownattheconclusionoftheprogram? ThereareahandfulofissuesregardingtheapplicabilitythatremaintobetestedatthispointintheESTCPprogram: Whatarethelimitsofgeologicinterferenceforsuccessfulclassification?WehavedocumentedtheeffectofmoderategeologyattheBealedemonstrationandhaveshownthattheseveregeologyatWaikoloamakesclassificationofsmallitemsverydifficult.Thebreakpointbetween“extracarerequired”and“toodifficulttobeeffective”isstilltobedeterminedprecisely. Siteswith20-mmprojectiles.Severaldemonstrationshaveshownthat20-mmprojectilescanbeclassifiedbutwehavenotyetconductedademonstrationatasite(otherthanastrafingrange)where20-mmprojectilesaremixedwithothermunitions.Becauseofthisweareunabletodocumentthedecreasedclassificationefficiencyduetothepresenceofsmallermunitionsitems.Asearchforagoodexamplesiteisunderway. Urbanandindustrialareas.Theclutterenvironmentinurbanandindustrialareasislikelytobequitedifferentfromthatonanisolatedrange.TheESTCPprogramhasnotdocumentedtheclassificationperformancepossibleintheseenvironments. OB/ODareas.Thesesitesarelikelytocontainmanypartialanddistortedmunitionsthatarestillhazardous.TheabilitytocorrectlyclassifytheseitemsusingthecurrentTOIlibraryhasnotbeendetermined.ESTCPhasademonstrationplannedatthePuebloArmyDepotthatshouldaddressthisissue. Ifasiteintheplanningstagepresentsoneofthesecharacteristics,itmaybewisetoperformatreatabilitystudyatthesite. BacktoTop Print StrategicEnvironmentalResearchandDevelopmentProgram(SERDP) EnvironmentalSecurityTechnologyCertificationProgram(ESTCP) Phone(571)372-6565 4800MarkCenterDrive, Suite16F16, Alexandria, VA 22350-3605 Media ContactUs CopyrightInformation Section508 Share