Mobile

Grave mapping in support of the search for missing persons in conflict contexts

Description
We review the current and potential uses of Geographic Information Software (GIS) and “spatial thinking” for understanding body disposal behaviour in times of mass fatalities, particularly armed conflict contexts. The review includes observations
Categories
Published
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  Gravemappingin   support   of    the   searchformissing   persons   in   con fl ictcontexts DerekCongram a, *,   MichaelKenyhercz b,c ,   ArthurGillGreen d a GlobalJusticeLab,MunkSchoolofGlobalAffairs,UniversityofToronto,315BloorSt.West,   Toronto,ONM5S     3K7,Canada b DepartmentofDefensePOW/MIAAccountingAgency,CentralIdenti  fi cationLaboratory,590MoffetStreet,BLDG4077,JointBase   PearlHarbor-Hickam,HI 96853,   USA c DepartmentofAnatomy,UniversityofPretoria,Private   Bag     x323,0007Arcadia,SouthAfrica d DepartmentofGeography,UniversityofBritishColumbia,1984WestMall,Vancouver,BCV6T1Z2,   Canada A   R    T   IC   L    E   IN   F   O Article   history:Available   online   27    July   2017 Keywords: Forensic   anthropologyForensic   archaeologySpatial   analysisGeographic   Information   ScienceForensic   Humanitarian   Action AB   S   T   R    A   C   T Wereviewthecurrentand   potentialusesof    GeographicInformationSoftware(GIS)and “ spatialthinking ” forunderstandingbodydisposalbehaviourintimesof    massfatalities,particularlyarmedcon fl ictcontexts.Thereviewincludesobservationsmadebytheauthorsduringthecourseof    theiracademicresearchandprofessionalconsultingontheuseof    spatial   analysisandGIStosupportHumanitarianForensicAction(HFA)tosearch   forthedead,theoreticalandstatisticalconsiderationsinmodellinggravesite   locations,and   suggestionson   howthisworkmaybeadvancedfurther.©2017ElsevierB.V.Allrightsreserved. 1.Introduction Rodrigo   Guerrero   Velasco,   the   mayor   of    Cali,   Colombia   from1992   to   1994   (re-elected   in   2011),   is   an   epidemiologist.   Asmayor   of acity   that   was   at   the   time   plagued   with   a   homicide   rate   of    124per100,000   residents,   he   adopted   an   approach   to   fi ghting   crime   thatthelocal   press   labelled   “ urban   acupuncture ”—   sticking   pins   in   amap   to   mark   crimes,   particularly   homicides.   This   “ hot   spot ” mapping   (which   isnow   routinely   digital   and   GIS-based)   allowedthe   municipal   authorities   to   focus   resources   on   the   “ sick ”   areas   of thecity   [1].Guerrero   Velasco   understood   the   value   of    visualisationandof    data-driven   inquiry.   During   his   tenure   as   mayor   and   for   twoyearsfollowing,   the   homicide   ratein   Cali   dropped   signi fi cantly   [2].One   of    several   maps   thatwere   exhibited   in   the   genocide   trial   of General   Ratko   Mladic   at   the   United   Nations   International   Tribunalforthe   former   Yugoslavia   highlighted   the   spatial   relationship   of schools   with   mass   execution   sites   (Fig.   1).Therelationship   mightseem   contrived   unless   you   know   that   up   to   8000   men   and   boysfrom   in   and   around   Srebrenica   were   detained   for   several   daysbefore   being   executed.   Because   most   of    the   buildings   in   the   areahad   limited   capacity,   the   detentions   were   mostly   in   schools,   anagricultural   warehouse,   and   a   cultural   centre.   Some   of    those   whowere   held   prisoner   and   survived   the   killings   were   able   totestifyabout   the   mass   executions   atand   around   these   detention   centres.Inmost   cases,   victim   bodies   were   transported   intrucksfromexecution   sites   to   nearby   burial   sites.   Understanding   the   spatialdynamics   and   logistics   of    detentions,   which   includes   knowing   theboundaries   of    the   area   under   control   of    those   responsible   for   thesubsequent   killings   (the   area   marked   “ RS ”   in   Fig.   1),   was   importantin   the   eventual   discovery   of    victim   burial   sites   [3].When   mass   fatalities   occur   duetonatural   disasters   orarmedcon fl ict,   of  fi cial   resources   for   interring   the   dead   and   investigatingthe   missing   often   become   overwhelmed,   forcing   improvisationaltreatment   of    both   statutory   and   customary   treatment   of    the   deadandmissing.   In   cases   including   illegal   killings,   burial   customs   maybedeliberately   violated   either   asa   means   of    concealing   evidence(i.e.,   victim   bodies)   of    crimes   or   asameans   of    disrespecting   thevictims   and   their   communities.   In   these   scenarios,   the   bodies   of    thedead   are   often   buried   anonymously,   transforming   them   into “ missing   persons ” .For   those   seeking   the   missing,   understandingsituational   variability   inburials   anddeviations   from   customaryandstatutory   burial   practices   is   paramount.   Knowing   the   circum-stances   of    disappearance   and   deathas   well   as   those   responsiblecanhelp   usdeduce   where   we   ought   to   be   looking   for   the   bodies   of thosewho   are   missing.Inthis   article,   we   emphasise   the   utility   of    spatial   thinking   andanalysis,   things   typically   eschewed   in   favour   of    oral   testimony   andwritten   documentation.   Spatial   analysis,   in   this   context,   involvesvisualising   an   area   of    investigation   and   assessing   spatial   relation-shipsamong   variables   that   in fl uence   how   and   where   bodies   areburied   (orotherwise   managed).   Weintroduce   some   GeographicInformation   Science   (GIScience)   tools   that   enable   more   effective *   Corresponding   author. E-mail   address:   derek.congram@utoronto.ca   (D.   Congram).http://dx.doi.org/10.1016/j.forsciint.2017.07.0210379-0738/©   2017   Elsevier   B.V.   All   rights   reserved.Forensic   Science   International   278   (2017)   260 – 268 Contents   listsavailable   at   ScienceDirect ForensicScience   International journal   homepage:   www.elsevier.com/locate/forsciint  investigation   of    the   missing,   presumed   dead.   Weillustrate   theseconcepts   and   methods   with   several   cases   from   our   appliedresearch.   The   aims   of    our   research   are   to:   (1)   supplementtraditional   investigative   efforts   and   (2)   explore   new   means   of investigation   using   GIScience.   More   than   simply   introducingconcepts   and   tools,   however,   we   advocate   spatial   analysis   as   amore   informed   wayof    preparing   for   disaster   tomitigate   the   social,psychological,   and   material   cost   of    not   knowing   the   whereaboutsof    those   who   havedisappeared   and   are   believed   to   havedied. 1.1.Background Typically,   those   who   investigate   missing   persons   cases   seek   outwitnesses.   Witnesses   describe   what   they   saw,   turning   memories, Fig.   1.   Mapsubmitted   as   evidence   in   the   trial   of    Ratko   Mladic,   Bosnian-Serb   General,   showing   territorial   boundaries,   schools   used   as   detention   sites   and   mass   execution   sites.Evidence   Reference   Number   0706-7941   in   Prosecutor   v.   Mladi  c,IT-09-82. D.   Congram   etal.    /    Forensic    Science   International    278(2017)    260 –  268   261  in   the   form   of    mental   images,   into   words,   which   are   documentedand   used   to   guide   investigations.   In   some   cases,   witness   state-ments   lead   investigators   to   a   speci fi c   place   (e.g.,   a   detention,execution,   or   burial   site)   and,   depending   on   the   perceivedreliability   of    the   information,   might   lead   to   an   excavation   insearch   of    a   grave   (a   prospection).   Ideally,   the   information   isaccurate   and   precise   so   thatbodies   can   be   found,   exhumed,identi fi edand   returned   to   family   for   culturally   appropriatetreatment.   Yet,   sometimes   the   information   provided   by   witnessesor   informants   is   not   reliable   and   no   burial   sites   are   located.   Newwitnesses   will   then   besought   in   the   hope   that   they,   in   turn,   will   beable   to   identify   a   burial   location.   This   methodological   loop   tends   tohavediminishing   returns   until   there   are   no   more   witnesses   withnew   information   and   investigators   simply   stop   searching   for   themissing.When   the   search   for   unmarked   burial   sites   fails,   we   seldomknow   what   went   wrong.   Did   the   witness   of    a   possibly   traumaticevent   simply   not   remember   accurately?   Was   their   informationgenerally   correct   but   imprecise?   For   example,   a   grave   prospectionmight   have   stopped   only   30   m   away   from   the   actual   burial   site.   Orperhaps   the   witness   was,   out   of    amisguided   desire   to   help,subconsciously   and   inaccurately   embellishing   the   facts.   Worse,   butplausibly,   a   witness   might   have   been   deliberately   deceivinginvestigators   by   providing   false   or   inaccurate   information.   Thoseofus   engaged   in   these   failed   searches   know   how   dishearteningthey   can   be.   Most   of    uscan   only   imagine   what   impact   failedprospections   have   on   the   families   who   are   seeking   the   missing.Driven   by   these   failures,   we   seek   to   develop   an   alternative   searchmethod.   One   that   does   not   exclusively   rely   on   witnesses   who   candraw   an   “ x ” on   amaptoindicate   a   burial   site   ordescribe   itslocation.   Instead,   we   examine   the   common   patterns   of    character-istics   of    the   known   locations   of    body   disposal   during   times   of    massfatalities   when   the   capacity   toregister   and   mark   the   burial   place   of thedead   is   overwhelmed   and   in   the   instance   of    criminaldisappearances   and   deaths.Since   the   early   2000s,   geospatial   technologies   have   becomeembedded   in   applied   research   in   awide-range   of    academicdisciplines,   innovative   business   models,   international   humanitari-an   organizations   and   civil   society   groups.   Geographic   InformationScience   (GIScience)   refers   toresearch   that   both   develops   and   usesgeospatial   technology   to   create   analytical   models   for   scienti fi cresearch.   Geospatial   technology   refers   to   a   wide   array   of    datagathering   instruments   that   produce   data   for   Geographic   Informa-tionSystems   (GIS).   Forexample,   Convergne   and   Snyder   discusshowgeospatial   technology   has   become   a   strategic   and   tactical   toolfor   United   Nations   peacekeeping   operations   [4].   Other   interna-tional   organizations   haveexplored   GIS   and   other   geospatialtechnology   for   mapping   inhumanitarian   contexts,   forhumanrightsmonitoring[5]and   insome   rare   instances   for   detectingsuspected   massgrave   sites   [6,7].Thenon-pro fi t Ushahidi   began   in2008   to   crowd-source   volunteered   geographic   information   (VGI)to   map   electoral   violence   in   Kenya.   In   2010   they   managedvolunteer   efforts   to   use   GIS   and   crowd-sourced   data   to   map   thestreets   in   Port-au-Prince   to   facilitate   the   delivery   of    humanitarianaid   in   post-earthquake   Haiti   [8].   Thewide   application   of    GIS   insuch   scenarios   belies   the   seemingly   short   next   step   tomappingvictims   of    such   disasters   (places   last   seen   alive   or   seen   dead,morgues,   hospitals,   and   body   disposal   sites).   Identifying   the   deadismore   easily   done   when   multiple   locations   can   be   linked   totriangulate   data   and   there   has   been   a   profusion   of    opengovernment   data   and   public   data   generation   by   way   of    almostubiquitous   mobile   phones   and   GPS-enabled   cameras.   Yet,   there   areconsiderable   procedural   issues   thatneed   to   beresolved   to   linksuchspatial   data   sets   to   the   expertise   and   spatial   analysis   requiredto   interpret   and   turn   the   con fl uence   of    data   into   actionableinformation.   We   are   now   light   years   beyond   pins   on   a   map,   andlinking   these   big   data,   open   data,   VGI,   and   otherdata   sets   totheexpertise   required   to   handle   and   interpret   them   could   lead   todramatic   changes   in   the   waythe   authorities   (orothers)   handle   thedead   in   times   of    disaster   and   war.GIS   is   aneffective   tool   for   visually   displaying   multiple   sources   of data   into   one   coherent   image,   often   depicting   latent   trends   notdirectly   observable   from   individual   data   sources.   Moreover,witness   testimony   does   not   need   to   be   abandoned   as   itcaninstead   be   coded   andincorporated   into   a   geospatial   database   thatmaintains   the   integrity   of    the   accounts   asattributes   of    spatiallocations.   Mapping   burial   sites   can   serve   multiple   purposes.Sometimes   exhumations   are   not   feasible   (e.g.,   ongoing   con fl ict,lack   of    resources   orexpertise),   orare   not   desired   (e.g.,   for   religiousreasons   or   political   sensitivity)   [9].   Simply   plotting   the   gravelocations   on   a   mapcan   be   critical   to   the   return   to   that   location   if conditions   change   in   the   future   (e.g.,   an   end   to   the   war   and   thedesire   for   humanitarian   exhumation,   identi fi cation,   and   repatria-tion).   In   addition   to   burial   sites,   mapping   places   of    disappearanceanddeath   allows   the   recording   of    related   geographic   features   andproxy   variables.   Proxy   variables   can   be   generated   through   spatialanalysis   of    information   either   from   available   datasets   orcarefullycoded   witness   testimony   (distance   from   road,   location   last   seenalive,polygons   of    known   battle   sites,   to   name   a   few).   In   thiscontext,   there   are   two   primary   purposes   of    mapping:   visualrepresentation   and   spatial   analysis.   Sometimes   analysis   is   simplyintuitive,   as   in   the   classic   crime   map   of    Guerrero   Velasco   in   Cali   toidentify   a   cluster   of    crimes   in   aparticular   neighbourhood   and   focustheir   investigative   efforts   on   that   place.   However,   GIScience   andspatial   statistics   enable   much   more   powerful   analysis. 2.   Materials   and   methods  2.1.   Geospatial   tools   and   data At   its   core,   aGIS   includes   a   sophisticated   database   that   allowsfor   the   acquisition,   management,   analysis,   anddisplay   of    spatialdata.   Data   within   a   GIS   is   categorised   as   either   vector   (points,   lines,andpolygons)   or   raster(anyimage   based   on   pixels,   such   as   a   .tiff    or.jpeg).   Vector   data   represents   discrete   features,   such   as   thecoordinate   point   of    agrave,   a   cemetery,   a   road,   or   a   territorialboundary.   A   raster   dataset   iscomprised   of    pixels;   in   thisinstance,pixelsrepresent   grids   of    varying   sizes   (resolution),   thathave   aspatial   component   and   anattribute   component   for   each   grid   cell.   Acommonly   employed   raster   dataset   is   a   digital   elevation   model(DEM)   in   which   the   spatial   attribute   values   for   each   cell   includeelevation   forthat   location.   A   more   familiar   raster   format   is   .jpeg,where   the   frame   isdivided   intomillions   of    cells   (measured   aspixels   persquare   inch)   and   the   attribute   value   for   each   cell   is   acolour,thus   composing   a   picture.   Unlike   traditional   databases,   theorganizing   principle   for   all   data   within   a   geospatial   database   isspatial   location.For   the   purposes   of    mapping   the   missing   and   deceased,   pointscan   beplotted   asvector   objects   into   a   geographic   reference   system(often   latitude   and   longitude)   that   is   linked   to   other   layers   such   aspolitical   boundaries.   Ideally,   each   point   corresponds   toanindividual   or   case.   For   each   individual ’ s   point,   several   non-spatialfeature   attributes   can   be   recorded   to   construct   a   database   of    themissing.   Attributes   can   include   information   such   as   fi eldidenti fi -cation   numbers   for   each   missing   person,   their   location   last   seenalive,political   af  fi liation,   civilian   status,   age,   sex,   nationality,stature,   and   identifying   markings   orcharacteristics   (such   astattoos,   dental   augmentation,   etc.).   Attribute   tables   can   beenteredinto   GIS   without   having   associated   coordinates.   This   is   particularlyimportant   in   the   event   of    a   grave   excavation   when   abody   isidenti fi edbecause   the   coordinates   can   then   be   added   totheattribute   table. 262   D.   Congram   etal.    /    Forensic    Science   International    278   (2017)    260 –  268  Geoprocessing   steps   such   asbuffering,   clipping,   and   measuringcanbe   conducted   on   selected   or   all   spatial   layers   (e.g.,   topographic,roadand   railway   networks,   land   use,   hospitals,   cemeteries,military   facilities,   morgues,   military   fi eld   maps)   and   entered   intoa   GIS   for   analysis.   Such   steps   allow   ustomeasure   the   relationshipsbetween   places   and   combine   data   sets.   De fi ning   all   possiblelocations   of    victim   remains   requires   pooling   different   types   of information   from   various   sources.   The   relevance   of    relatinglocations   from   different   types   of    spatial   data   layers   can   beseenin   the   investigation   of    Malaysia   Airlines   MH17.   The   commercialaircraft   carrying   282   passengers   is   believed   to   have   been   shotdown   over   the   Ukraine   in2014.   Bodies   recovered   atthe   crash   sitewere   moved   via   train   by   rebel   authorities   to   a   nearby   town   [10].The   spatial   relationships   of    the   crash   site,   rivers,   the   politicalboundaries   and   military-controlled   territory   within   which   thecrash   occurred,   a   missile   launcher   position,   and   the   road   and   trainnetworks   all   in fl uenced   how   victim   bodies   were   handled   followingimpact   (at   multiple   locations   because   of    the   mid-air   explosion   andfragmentation).Theintroduction   of    spatial   data   into   geodatabases   can   alsoenhance   our   understanding   of    errors   in   data   collection   andprocessing   that   might   escape   attention   in   conventional   non-spatial   databases.   In   the   former   Yugoslavia,   particularly   in   Bosnia – Herzegovina   from   1996   to   2001   and   Kosovo   in   1999   and   2000,multiple   investigative   entities   operated   concurrently   [11].Theseorganisations   used   different   standards   of    recording   information,whichcaused   problems   coordinating   knowledge   and   action   on   theground.   Gathering   data   on   the   deaths   at   Srebrenica   in   Bosnia   was   aparticular   focus   of    the   Of  fi ce   of    the   Prosecutor   (OTP)   of    theInternational   Criminal   Tribunal   for   the   former   Yugoslavia,   whichdedicated   resources   and   expertise   to   document   killings   in   a   verythorough   manner.   Theprimary   interest   of    the   OTP   was   toinvestigategrave   violations   of    international   criminal   law,   includingcrimes   against   humanity   and   genocide.   Both   of    these   crimes   aredemonstrated   by   systematic   killings,   whichcan   require   viewingvictim   graves   at   a   smaller   (i.e.,   “ zoomed   out ” )   scale,   rather   than   onan   individual   basis.   Other,   smaller   (in   personnel   and   budget)organisations   recovering   bodies   in   the   Balkans   generally   operatedona   site-by-site   basis   and   their   goalswere   primarily   humanitarian,soless   concerned   with   reconstructing   a   larger   temporal   andgeographic   narrative   as   would   beimportant   to   acriminalinvestigation   of    genocide.   As   such,   there   is   farless   detail   availablefrom   their   work   thatcan   now   beused   to   analyse   patterns   and   scaleof    deaths   and   burial   during   and   following   the   wars.   If    we   rely   onthis   available   geographic   data,   the   more   thorough   andnumerouscasesdocumented   in   Srebrenica   can   constitute   a   sampling   error   formodel   design   and   in   the   actual   analysis   of    data   (i.e.,   the   sample   willbeskewed).There   are   several   other   data   entry   inconsistencies   that   cancause   errors   in   analysis.   For   example,   places   might   havenames   thatarespelled   differently   according   to   the   language   being   spoken   orwritten   during   the   documentation   process   (e.g.,   Kosovo   or   Kosova,Table   1).   By   giving   spatial   coordinates   priority,   such   differentialnames   are   resolved   in   a   spatial-enabled   geodatabase.A   second   data   problem   relates   to   quality.   Although   there   aremany   free   or   inexpensive   digital   maps   available   from   onlinerepositories,   governments   orother   organisations,   the   precisionand   accuracy   of    them   can   differ   greatly.Fig.   2shows   a   map   withroad   networks   from   three   different   sources.   One   source   (withroads   drawn   using   aburgundy   line)   is   much   more   detailed   than   theothers,   but   the   others   do   mark   some   roads   thatthe   fi rst   does   not.These   maps   can   be   merged,   but   it   can   take   time   and   thecompleteness   of    the   fi nal   map   might   still   be   lacking.Another   bene fi tof    using   geospatial   tools   such   as   GIS   isthat   theycanbe   deployed   atseveral   scales.   GIS   can   support   spatial   analysisincluding   bone   microstructure   [12],   mapping   body   positionswithin   a   mass   grave   to   assist   with   commingling   resolution   [13],mapping   deposits   of    bodies   within   a   grave   [14],mapping   asitewithin   a   geographical   context   and   mapping   sites   relative   to   eachother   [15],   as   well   as   graves   relative   to   other   points   of    interest   suchas   schools,   as   illustrated   above.Ultimately,   exploring   spatial   relationships   within   particulararmed   con fl ict   contexts   gives   amore   quanti fi able   understanding   of the   dynamics   that   caused   people   to   gomissing.   While   this   can   help fi ndthe   remains   of    thosewho   havedied,   thismay   also   help   analyseand   understand   comparisons   of    spatial   variation   and   relationshipsof    similar   events   among   distinct   countries.   A   greaterunderstand-ingof    the   variation   in   human   behaviour   regarding   body   disposalwill   help   to   develop   theories   thatguide   further   analysis   andimprove   the   effectiveness   of    searches   for   the   missing,   identi fi ca-tion   of    remains,   and   return   of    the   missing   to   their   families   fordigni fi ed   burial.  2.2.   Models    for    locating    the   missing  A   fundamental   concept   in   modelling   human   spatial   behaviouris   that   humans   interact   with   their   environments   in   patterned,   non-random   ways[16 – 22].   The   ways   in   which   humans   have   predictablyexploited   their   environments   has   been   the   basis   of    archaeologicalsiteprediction   that   hasbeen   employed   byCultural   ResourceManagement   fi rms   for   decades   [23].   Congram   and   Kenyhercz   haveargued   that   human   behavioural   ecology,   speci fi cally   optimalforagingtheory   (OFT),   can   beused   as   a   theoretical   frameworkforapplying   site   prediction   modelling   to   aidin   locating   the   missing[24].   Brie fl y,OFT   hypothesises   that   natural   selection   ispreferentialto   animals   whose   behavioural   strategy   maximises   their   energyintake   perunit   of    time   spent   foraging   for   resources.   Congram   andKenyherczextended   OFT   to   understand   the   nature   of    clandestinebody   disposal   by   positing   thatclandestine   burial   location   is   afunction   of    time   spent   with   remains   (analogous   toforaging   time)and   avoiding   detection   (selection),   which   is   bounded   byeitherculture,   local   environments,   or   both   [24].As   mentioned   above,   there   are   many   spatial   variables   availabletomodel   the   location   of    potential   sites,   however,   few,   if    any,   aregoingto   be   explicitly   related   to   the   death   event   orsubsequentburial   activity.   Instead,   spatial   variables   are   oftenused   asproxiesfor   human   behaviour,   orcognitive   decision-making.   Whitleyidenti fi edthree   different   classes   of    proxy   variables   thatinformcognitivedecision-making:   (1)   direct   causal   reference;   (2)   indirectcausal   reference;   and   (3)   non-causal   reference   [25].   Direct   causalreferences   tie   an   environmental,   or   spatial,   variable   to   some   sort   of cognitive   behaviour.   For   example,   a   viewshed   isa   map   created   withGIS   that   showsthe   area   that   isvisible   from   a   fi xed   location.   Thevisiblearea   can   directly   impact   decision-making,   which,   in   thecurrent   context   might   relate   to   avoiding   detection   while   disposingof    remains.   Indirect   causal   references   do   not   explicitly   in fl uencecognition,   but   will   affect   the   waythat   latent   variables   mayin fl uence   decision-making.   To   illustrate,   distance   maps,   or   cost-distance   maps   (distance   maps   that   are   bounded   by   other   variablessuch   asslope,   natural   or   cultural   barriers,   etc.)   could   be   used   as   a  Table   1 Distinct   place   names   for   the   same   districts   according   tolanguage   inKosovo/Kosova.Albanian   Serbo-CroatDrenas   GlogovacFerizaj   Uro š evacFushë   Kosovë   Kosovo   PoljeKamenicë   Kosovska   KamenicaRahovec   OrahovacSkënderaj   Srbica D.   Congram   etal.    /    Forensic    Science   International    278(2017)    260 –  268   263  proxy   for   familiarity   with   a   region,   whichassumes   thatpeople   aremore   familiar   with   their   immediate   vicinity   than   they   are   of regions   further   away,   thusunconsciously   bounding   their   potentialdisposal   sites.   Lastly,   non-causal   references   use   proxies   for   somesort   of    behaviour   even   though   there   is   no   direct   relationship.   Takethe   viewshed   example   again.   The   line   of    sight   from   aparticularvantage   has   a   direct   relationship   to   the   individual   from   thatvantage.   However,   the   vantage   point   might   also   allow   an   individualto   hear   better   or   worse   from   a   location,   which   cannot   bedirectlymeasured   in   GIS.Predictive   modelling   of    site   locations   is   both   an   inductive   anddeductive   process.   The   inductive   aspect   starts   with   compiling   adatabase   of    known   site   locations   and   plotting   them   in   a   commoncoordinate   system.   Simply   displaying   spatial   data   layers   in   aninteractive   GIS   facilitates   the   identi fi cation   of    spatial   patternsusing   the   human   eyeand   expert   knowledge   of    events   andgeographic   areas,    just   as   Velasco   used   pins   on   a   paper   map   toelucidate   homicide   clusters   in   Cali.   Spatial   data   analysis   recognisessome   unique   barriers   to   analyses   and   accounts   for   problems   suchas   the   Modi fi able   Areal   Unit   Problem   (MAUP).   TheMAUP   occurswhen   point-based   data   are   aggregated   intonew   levels   of    analysis(often   polygons)   as   is   often   done   when   district   reports   of    criminalevents   are   aggregated   into   table   formats.   The   aggregation   of    points(such   as   burglaries)   and   subsequent   averaging   of    attribute   data(such   as   cost   of    items   stolen)   obfuscates   and   leads   to   aloss   of understanding   of    spatial   patterns   of    behaviour,   especially   whenclusters   of    events   may   cross   over   and   bedivided   intoneighbouringareas   (such   as   police   districts).   In   other   words,   burglaries   thatstraddle   a   boundary   might   not   beseen   as   being   related   if    they   areonly   looked   at   within   individual   police   district   boundaries.   Thespatial   display   of    data   in   GIS,   even   without   statistical   modelling,allows   amore   effective   visualization   and   analysis   of    bothinductiveand   expert   knowledge.   Identi fi ed   patterns   might   then   beinvestigated   through   fi eldwork   or   through   statistical   modelling.Statistically   signi fi cant   clusters   can   be   identi fi ed   through   avariety   of    analyses   including   Getis-Ord   Gi   (Hot   Spot)   analysis,Ripley ’ s   K-function   cluster   analysis,   and   Average   Nearest   Neigh-bour   tests.   These   analyses   examine   spatio-temporal   relationshipsof    event   observations   as   well   as   the   spatio-temporal   clustering   of observation   attributes   (for   example,   crime   rates   or   event   magni-tude)   for   statistically   signi fi cant   clusters.   After   signi fi cantclustershave   beenidenti fi ed,   the   spatial   relationship   between   environ-mental   and   cultural   variables   can   betested:   distance-to-road   (orwater,or   railways,   battle fi elds,   site-last-seen,   etc.),   slope,   view-sheds,   elevation,   surface   geology,   land-use,   population   density,   anddemographic   and   economic   distribution   maps   to   name   a   few.UsingGIS,   variables   can   be   assigned   to   each   of    the   signi fi cantlyclustered   points.   Further,   to   testfor   signi fi cance,   one   can   comparethe   variable   values   at   each   known   sitelocation   to   those   at   randomsitelocations.   This   will   show   if    the   environmental   and   culturalvariables   show   any   signi fi cant   pattern   particular   to   the   knowngrave   sites   when   compared   to   a   random   distribution   of    placeswhere   there   are   no   sites.   Using   GIS,   it   ispossible   toset   a   study   areaboundary   and   populate   itwith   a   random   distribution   of    points. Fig.   2.   Three   road   network   maps   for   the   same   location,   each   differing   in   precision   and   completeness.   One   public   source   maps   ( “ OSM ” ,   which   stands   for   OpenStreetMap,above)is   more   complete   andprecise   than   others   from   privatesources,   but   each   of    the   three   has   roads   marked   that   are   absent   in   theothers.   Credit:  © OpenStreetMapcontributors,   data   available   under   Open   Database   License.264 D.   Congram   etal.    /    Forensic    Science   International    278   (2017)    260 –  268
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x