Scribd
Upload
169 views

Data Visualization Workshop Report

This document summarizes a workshop on data visualization to support ecosystem-based management that was held in February 2013 at the Gulf of Maine Research Institute in Portland, Maine. The workshop brought together participants from NOAA units, partner agencies, and academic institutions to share information and develop synergies around data visualization tools to help communicate ecosystem science and support ecosystem-based management decisions. Presentations at the workshop covered a wide range of topics, including visualizing complex ecosystem data and models, designing visualizations for different audiences, and tools for accessing and analyzing various types of environmental data. The goal was to facilitate exchange of information around existing visualization capabilities and needs to better inform management and the public about ecosystem science.

Uploaded by

mickamal
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
169 views

Data Visualization Workshop Report

This document summarizes a workshop on data visualization to support ecosystem-based management that was held in February 2013 at the Gulf of Maine Research Institute in Portland, Maine. The workshop brought together participants from NOAA units, partner agencies, and academic institutions to share information and develop synergies around data visualization tools to help communicate ecosystem science and support ecosystem-based management decisions. Presentations at the workshop covered a wide range of topics, including visualizing complex ecosystem data and models, designing visualizations for different audiences, and tools for accessing and analyzing various types of environmental data. The goal was to facilitate exchange of information around existing visualization capabilities and needs to better inform management and the public about ecosystem science.

Uploaded by

mickamal
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 56

WorkshoponDataVisualizationtoSupport

EcosystemBasedManagement

NorthAtlanticRegionalTeam,NART
GulfofMaineResearchInstitute
Portland,Maine
February56,2013

Contents
Introduction....................................................................................................................................3
Presentations..................................................................................................................................5
Changingperspectivesaboutachangingecosystem.................................................................5
Visualizingcomplexity:howcanweviewfoodwebandmultispeciesdataanddynamic
ecosystemmodeloutputs..........................................................................................................7
VisualizationandGamingofEcosystemModelScenarios.......................................................11
Gettingmorefromdatathanjustwherethefishare..............................................................13
Visualizinghighlyspatialandtemporaldata..............................................................................2
Challenges,lessonslearned,andaFEWpotentialvisualizationsolutionsforsynthesizing
sciencetoinformManagementinsouthFlorida.......................................................................4
Dynamicconceptualizationofhabitat........................................................................................6
Designinggeospatialdatavisualizationsforageneralaudience...............................................7
Dealingwithscientificandcommunicationuncertaintythroughdatavisualization...............12
NEFSCDataServer:access&visualizationwithPython...........................................................13
Webaccess,analysisandvisualizationofstandardizedoceanographicandmeteorological
data...........................................................................................................................................15
OceanographicdataanddataservicesatPacIOOS,aregionalassociationofIOOS...............17
Dataandvisualizationintegrationviawebbasedresources...................................................19
VisualizingFitnessforPurpose.................................................................................................21
VisualizationsofHabitatandSitingDecisions..........................................................................23
Habitatmodelingandvisualization..........................................................................................25
Communicatinganunderstandingoflowertrophiclevelsinmarineecosystems..................27
FindAndVisualizeNOAAData:AnOverviewoftheNOAAEarthInformationSystem(NEIS)
andTerraViz..............................................................................................................................28
Gettingtoobservingdatausigntheclimatologytoolkit..........................................................29
Developingaclimatedatavisualizationtoolformarineecosystemmanagers.......................31
ContributionsviaCorrespondence...............................................................................................36
Recommendations........................................................................................................................37
Principles:..................................................................................................................................37
Actions:.....................................................................................................................................38
Appendix1.Agenda......................................................................................................................39
Appendix2.Participants...............................................................................................................40
Appendix2.Participantscontinued..........................................................................................41
Appendix3.VisualizationProductsandServices.........................................................................42

CoverartbyK.Friedland

Introduction

Artistsneedtoputasmuchtime,energy,andresourcesintomarketingtheirworkastheyput
intocreatingit.Ifyouspend10hourscreatingyourartwork,youneedtomatchthatwith10
hoursofpromotionandmarketingofyourart.

KennethProudfoot

AsNOAAmovestoanecosystembasedmanagementapproach,itwillbeimportanttoensure
anunderstandingofthescienceavailable,itsapplicationtomanagement,andhowtobest
communicatethistothepublic.NOAAofficeshavevaryingcoreresponsibilitiesthathaveled
tothedevelopmentofadiverserangeofdatavisualizationtoolswithintheagency.Manyof
thesetoolshavethepotentialtohelpsolvethedatavisualizationproblemsofotherNOAA
offices;however,researchersandmanagersoftendonothaveaworkingknowledgeofdata
visualizationcapabilitiesdevelopedoutsidetheirownoffice.NOAAalsohasagencyand
academicpartnersthatcanprovidevisualizationcapabilitiestoapplytothesharedgoalofa
betterunderstandingofecosystemstructureandfunction.
WhatisthelevelofinvestmentintransmittingdataeffectivelythatNOAAshouldmake?Asthe
quotationabovesuggestsforartists,aconcomitantamountoftimepromotingandmarketing
artshouldmatchtheinitialinvestmentinthecreativeprocess.Theanalogycanbeappliedto
science,andecosystemmanagementinparticular,thatscientistsshouldconsiderinvesting
moretimeandresourcestoinsuretheirfindingsareeffectivelycommunicatedtoother
scientists,managers,stakeholders,andthepublic.
Whatcontinuestobeanareaofrelativedatavisualizationweaknessisthepresentationofthe
sciencesupportinganecosystembasedapproachtomanagement.NOAAhastremendous
fisheriesandenvironmentaldataholdings,yetinsomeareasthedataareonlysharedamong
scientists,managersandstakeholderswiththeuseoftablesandsimplisticgraphs.Itis
importantthatNOAAimproveitsunderstandingoftheavailableecosysteminformationin
ordertobetterinformresourcemanagementdecisions.Forexample,fisherymanagement
councilsandthefishingindustryarebeingaskedtoincorporateecosystemconsiderationsand
toadoptanecosystemsapproachtomanagement,yetmanymembersofthiscommunityare
stilldevelopinganunderstatingofecosystemconceptsandthemodelsbeingusedforthe
provisionofmanagementadvice.Managementscenariotestingandcomparisonunder
conventionalstockmanagementhasbeenrelativelytractable,typicallytakingtheformof
bracketedcomparisonsinvolvingalimitedsetofoptions.However,scenariotestingunderan
3

ecosystemapproachwillbemorechallengingtoconductandexplain.Therangeanddepthof
datafromecosystemsimulationsshouldbepresentedinawaythatreflectstheincreased
informationcontentofthesemodels.Itcouldbeenvisionedthatmanagementcouncilsmight
meetinavisualizationcapablemeetingroomtoconsidermodelsandscenarios,wheretesting
isconductedinteractivelywiththeparticipantsseeingnotonlythefisheryresponse,butalso
thewayclimatedriversareaffectingthephysicalaspectsoftheecosystemandhowalllevelsof
thebiologicalcommunityareresponding.ThebiologicalcommunitiesthatNOAAneedsto
considerinanecosystemmannerforthefisherymanagementprocessandothermanagement
processesarebroad(e.g.protectedresources,habitat,andclimatedrivers).Ecosystem
scientistsacrosstheagencyhavearesponsibilitytobetterinformthepublicaboutecosystem
functionandchange;bettervisualizationtoolswouldalsoaddressthisissue.
ThegoalofthisworkshopwastofacilitateinformationexchangebetweenNOAAunits,
cooperatingagenciesandacademicinstitutionsthathavedatavisualizationcapabilitiesand/or
requirementstoinstituteecosystembasedapproachestomanagement.Theworkshop
providedanopportunityforpractitionerstoshareinformationanddevelopsynergiesto
addressthegeneralproblemofdatavisualizationtosupportecosystembasedmanagement.
TheworkshopwasheldFebruary56,2013attheGulfofMaineResearchInstitute,Portland,
Maine.TheagendaforthemeetingcanbefoundinAppendix1.Theworkshopwaschairedby
K.Friedland,SamuelChavezservedasrapporteur,andafulllistofparticipants,includingthose
joiningthemeetingremotely,canbefoundinAppendix2.Theworkshopparticipantsalso
collatedatableofVisualizationProductsandServiceswhichcanbefoundinAppendix3.
Throughoutthetext,figurenumbersareinreferencetothesectioninwhichtheyappear.
Thisreportisasummaryoftheactivitiesoftheworkshopprovidedbytheworkshop
participantstotheNART;thereporthasnopeerreviewstatusandisnotpartofanyreport
series.

Presentations
Changing perspectives about a changing ecosystem
KevinFriedland,NationalMarineFisheriesService,Narragansett
TheNortheastShelfEcosystemisexperiencingdramaticchangeinitsphysicalenvironmentand
theresponseofthebioticcommunity.Thispastyear,2012,wasthewarmestyearever
recordedfortheecosystem.Longandshorttermtimeseriesarebeingusedtoexpressthe
trendsinecosystemtemperature.Longtermtrendsarebasedontemperaturerecordsdating
backtothe1800sandshowthatthedramaticincreaseintemperaturein2012exceeded
previoushightemperaturesrecordedinthe1940s(seefigure1).Itwassuggestedbythe
workshopthattemperaturefortheecosystem

Figure1.LongtermseasurfacetemperaturefortheNortheastShelfecosystemderivedfromtheERSST
database.

shouldbeshownincontexttothechangesintemperatureonaglobalscale.Theseglobal
analysesshowthattheNortheastShelfisoneofthemostrapidlywarmingecosystemsinthe
world,suggestingtheecosystemposesuniquemanagementchallenges.Mostcomprehensive
depictionsofecosystemtemperaturearebasedonseasurfacetemperaturesinceithasbeen
measuredmorefrequentlyoverthelongtermandisreadilyavailablefromsatellitesensors.
However,recentlydevelopedmodelhindcastsmayprovidesourcedatatodevelopsimilarly
comprehensivegriddedbottomtemperaturedataproductstocomplementtheseasurface
dataproducts.Thesedataresourcesshouldbedevelopedforcomplementaryusewithexisting
observationaldataresources.WiththereleaseofnewIPCCAR5assessmentdata,the
communityshouldbemadeawareoftheprojectedtrendsforwatertemperatureinthe
ecosystem.
5

Thebaseofthefoodchainisdrivenbyplanktonproductionthatispatternedindistinctseason
bloomcycles.Withthechangesintemperatureandsalinityoccurringintheecosystem,there
havebeenchangesrecordedinbothprimaryproductionandintheabundanceandspecies
compositionofzooplankton,whichformthebaseofthefoodchainforresourcespecies.
Temperatureandsalinitywillhaveadirecteffectonproductionthroughphysiologicaleffects
onlowertrophiclevelorganismsandthroughotherhabitatalterationslikechangeinthe
stratificationofthewatercolumn,whichcanaffectbloomtiminganddimensions.Inrecent
years,thespringbloomontheNortheastShelfhasstatedearlier(seefigure2).Theblooms
havealsobeenassociatedwithhigherwatercolumnplanktonbiomassasindicatedby
chlorophyll

Figure2.SpringbloomchlorophyllconcentrationandstartdayfortheNortheastShelfecosystemusingdata
fromSeaWiFSandMODISoceancolorsensors.

concentrations.Thismayimpactotherspeciesinavarietyofwayssuchaschangingfeeding
opportunitiesforearlylifehistorystagesorimpactingtheflowofenergyintheecosystem.How
thisinformationisperceivedbyabroaderaudiencewasdiscussed.Itwasagreedthat
supportingillustrationsofhowphysicalfactorscombinetoshapebloomswouldbeuseful.
Thesesameillustrationsshouldalsodrawattentiontothelinkagesbetweentheproductionof
loweranduppertrophiclevels.
Awelldocumentedconsequenceofrecentchangeintheecosystemhasbeenchangeinthe
distributionofspecies.Speciesdistributionhasbeenreportedintheresearchliteratureandis
describedinanumberofwebresourcesthatgenerallysupplypointofoccurrencedata.The
workshopdiscussedanumberofideasforbothaframeworktoadvanceregionaldescriptionof
speciesdistributionalandsomeofthedataelementsthatmightbeusefultoprovidetoboth
researchandmanagementsectors.Thereareanumberofdataproductsthatprovide
distributionpointofoccurrenceofspecies,itwasagreedthatstatisticaldepictionsof
continuousdistributionwouldbeusefulonanumberoflevels.And,characterizationof
distributioncouldbeenhancedwithmeasuresofcenterofmassandkernaldensitydata.
6

Visualizationtoolstorelatespeciesdistributiontootherparametersliketemperaturewouldbe
desirable,aswouldanimateddepictionsofdistribution.Finally,themorecomprehensive
developmentofnichespacemodelsforawiderangeofspeciesshouldbeencouragedsothat
futuredistributionscanbeprojectedwiththedatafromclimateprojectionmodels.The
workshoprecommendedtheformationofaworkinggrouponvisualizingspeciesdistribution
(seerecommendations).


Visualizing complexity: how can we view food web and multispecies data and
dynamic ecosystem model outputs
SarahGaichas,NationalMarineFisheriesService,WoodsHole
Foodwebsshowtheenergeticrelationshipsbetweenmanyspeciesorcomponentswithinan
ecosystem.Thisinformationcanbeusefulwhenconsideringmanagementoptionsforhuman
impactsoninteractingspeciesorontheecosystemasawhole;however,foodwebscanbe
incrediblycomplexinlargemarineecosystemswherehundredsofspecieshavethousandsof
interactions,manyofthemnotdirectlyobserved.Therefore,muchofourunderstandingof
marinefoodwebscomesfromorganizingavailabledataintovariousmodelframeworks.
Modelsoffoodwebsandmultispeciessystemscanbegenerallydividedintostaticanddynamic
types.Staticfoodwebmodelsshowtherelationshipsbetweencomponentsatasnapshotin
time,whiledynamicmodels(sometimesbasedonthestaticmodels)showhowtherelated
specieschangeovertime.Bothtypesofmodelcanbeimplementedatanyspatialscale.The
purposeofthispresentationwastoshowarangeofexistingvisualizationsoffoodwebdata
andmodeloutput,togetfeedbackonwhichvisualizationsmightbemostusefulforwhich
targetaudiencesandpurposes.Further,thispresentationwasintendedtoprovokediscussion
ordevelopmentofimprovedvisualizationsthat1)adequatelycapturecomplexitywithout
overwhelmingthetargetaudience;2)arevisuallyappealingwhilepreservingscientificcontent;
and3)communicateanappropriatelevelofuncertaintyinobservationaldataormodel
outputs.
Visualizationsofstaticfoodwebs(Fig1)canrangefromsimpledrawingsofpredatorsandprey
connectedbyarrows(1a)tohighlycomplexboxdiagrams(1b),engineeringdiagramswith
numericflows(1c),andabstractnetworkswithnodesandlinkscodedforinteractionstrengths
andflowrelationships(1d).Theleastcomplexfoodwebsarethosewiththefewest
compartmentsandrelationships;reductionofcompartmentscanbeachievedbyeither
focusingonthefoodwebsupportingasinglespecies(Fig1a)orbyaggregatingmanyspecies
intofewer,moregeneralizedfunctionalgroups(Fig1c).However,therearetimeswhen
complexityisthepointofthevisualization.Also,withinteractivevisualizationcapability,users
7

canworkthoughthecomplexityofafoodwebsuchasFigs1bor1dbyselectingindividual
speciestoseetheirrelationshipswithinthelargercontextofthefullfoodweb.The
visualizationsoftwareusedtoproducebothFigs1band1dhasthisinteractivecapability,
althoughonlythenetworksoftware(Gephi,http://gephi.org/)iswidelyavailable.
Visualizationsofdynamicecosystemmodeloutputs(Fig2)cansimilarlyrangefromsimpletime
seriesplotsofmultipleattributesforasinglespecies(2a)orthebiomasstrendofanaggregated
groupfromthefoodweb(2b)throughmorecomplexmultispeciestimeseriesplotsor
aggregatespeciesresultsoverarangeofmodeledconditions.Uncertaintyinthemodeloutputs
canbeexpressedasenvelopesaroundresults(2c),variationsofboxandwhiskerplots(2d),or
ordinationsofresultsoverawiderangeofparametervaluesandmodeledconditions(2e).
Thesevisualizationsrequireconsiderablymoreexplanationtoorientviewersthanthestatic
foodwebvisualizations,andarelikelytofindamorereceptiveaudienceamongscientiststhan
amongthegeneralpublicintheircurrentform.However,theresultsofdynamicmodelsare
likelytobeusefulinamanagementcontextwheremanagersneedtovisualizethepotential
impactsofclimate,speciesinteractions,orhumanactivitiesoncomplexsystemsovertime.
Here,amoreinteractiveformatmayalsobehelpful,aswellascreativeuseofanimationto
showchangingsystemattributesovertime.Thechallengesofreflectingappropriate
complexityanduncertaintyindynamicmultispeciesmodelvisualizationsareconsiderable,but
maybeovercomewithacombinationofiterativeworkwithtargetaudiencesandcreativeuse
ofinteractivemedia.

Figure1.Staticfoodwebvisualizations.


Figure2.Dynamicfoodwebmodelvisualizations,without(top)andwith(bottom)uncertainty.

10

Visualization and Gaming of Ecosystem Model Scenarios


RobertGamble,NationalMarineFisheriesService,WoodsHole
Visualizationoftheoutputsfromecosystemmodelsisnotaneasytask.Therewillbemany
scenariosrun,withmultipleoutputsofimportance.Specifically,managerswillwanttoknow
thingslike:biomassoftargetspecies(ecologicallyorcommerciallyimportantspeciesaswellas
protectedgroups),economicimpacts,unanticipatedimpactsontheecosystemfrom
managementactions,andmanyotherpossibilities.Thisreinforcestheneedtohaveclear
managementobjectivesbeforerunninganymodel,butthepresentationoftheresultsstillwill
notbetrivial.Oneexampleofoutputspotentiallyusefultomanagersistheproportionof
impactsfromfishing,environmental,andspeciesinteractionsandhowthosemightchange
underdifferentscenarios(Figure1).Inthefigurebelow,allgroundfishinamultispeciessurplus
productionmodelhadtheirgrowthratereducedby10%inaclimatescenario.Themodelwas
designedtoseparateouttheeffectsonbiomassfromfishing,climate,andspeciesinteractions.
Thefigureshowsthatresultsinevenrelativelysimplescenarioscanbecomplexand
unintuitive.InspiteofYellowtailflounderbeingincludedinthegroupwitha10%reductionof
growthrate,itincreasedinbiomassbytheendoftherunduetocompetitivereleasebyother
membersofthegroundfishgroup.

Figure1.Changestobiomassovertherunofamultispeciesproductionmodelduetodifferentpressures.Values
belowthe0lineindicateanincreaseinbiomassduetoapressure,valuesabovethe0lineindicateadecrease.

Often,whatwillbemostimportanttomanagerswhenlookingatresultsfromecosystem
models,isacomparisonoftradeoffsinvariousoutputscomparedtoobjectives.Forexample,
onecouldbalancebiomassofcommerciallyimportantspeciesagainstmarinemammal
protectionaswellasforagefishwhichsupportbothgroups.Aradarorspiderplotisoneway
todisplaysuchdata,andtheexamplegivenisfromamultispeciesproductionmodelinwhich
11

foragefisharefishedatdifferentlevels.Theplotshowstheresponseofratiosofbiomasstoa
targetvaluefor6differentobjectives,wherethetargetvalueisrepresentedbythe1.0lineon
theplot(Figure2).OthervisualizationenvironmentssuchasGoogleEarthorTerraVizshould
alsobeexploredinthecontextofecosystemmodelswithageospatialcomponent.

Figure2:Aradar/spiderplotshowingtheeffectsofdifferentlevelsofharvestonforagefishcomparedto6
targetreferencepoints(representedbythe1.0redline).

Itwillalsobeofgreatusetomanagersandtobeabletointeractivelyexplorescenarios.These
canrangefrommorecomplexGUIinterfacesthathidetheunderlyingmodels(althoughthe
equationsshouldalwaysbemadeavailableinsomefashion),towebbasedinterfaceswhich
allowpickingspecificsetsofparameterstoviewthechangesinanumberoftypesof
visualizationsillustratingtradeoffs,toevensimplerexploratorymodels.Thegaming
environmentcouldbedynamicwithamodelbeingrunbasedonthechangedparameters,or
staticwiththemodelrunsalreadyhavingbeendoneinadvance(likelynecessaryforsomeof
themorecomplexecosystemmodels).Exploratorymodelscouldbeassimpleaswhiteboards,
orotherhandsongamessuchasusingmultipledecksofcardstorepresentimportant
mechanismsinanecosystemmodel.Whilenomanagementdecisionswouldbemadefrom
12

suchsimplegamingenvironments,theycouldprovideanunderstandingofwhatthemost
importantfeaturesofanecosystemmaybe.


Getting more from data than just where the fish are
ScottGeis,NationalMarineFisheriesService,WoodsHole
ArcGISisasystemdesignedtodisplay,integrate,andsynthesizegeographicanddescriptive
informationfromvarioussources.Thisplatformallowsscientiststopresentcomplex
informationbycombiningimageryanddatasetsfrommultiplesourcesintovisuallyappealing
maps.ThemyriadoftoolsprovidedbyArcGISallowforstatisticalanalysistobeperformed,and
recentadvancementsinArc10.1allowforenhancedevaluationandvisualizationoftimeseries
data.CommonapproachestotimeseriesdatavisualizationwithGeographicInformation
Systems(GIS)havebeenmetwithcriticismbecausetheyattempttodisplaydatawiththree
dimensionalcharacteristicsonatwodimensionalmap.Sincedatahasbothaspatialanda
temporalcontext,responsetothesecriticismshasbeentomovetowardsanalysesthat
rendereddataonathreedimensionalsurfaceandashifttotoolsthatperformspacetime
analyses.Intheend,noonetoolprovidesaonestopshottosuccessfullyvisualizingtimeseries
data.Insteaditisacombinationoftoolsandtechniquesthatcanproducethebestresults.
CommonApproachestoTimeSeriesVisualization:
WhilethevisualizationoffisheriesdatainGISisanestablishedpractice,commonapproaches
totheextensionofGIStotimeseriesdatahasbeenmetwithmixedreviews.Common
approachestoanalyzingdatawithspatialandtemporaltrendshavefocusedontheuseof
densitytoolstoproduceheatmaps(Figure1),oranimationstobreakthedataupintoa
seriesofsnapshotsdisplayingchangesinsomevariableovertime(Figure2).Whilebothof
thesemethodsprovideagoodmethodforexploringpatternsinthedata,theycontain
limitationswhenrepresentingthreedimensionaldata(xandylocation,plustime)withatwo
dimensionalmap.

13


Figure1:HeatMapexample

Figure2:Seriesofsnapshotsfromatimeseriesanimation

Heatmapsarebeneficialforrepresentingthegeographicdensityofpointfeatures.Theyare
ofteneasytoremember,andtypicallyutilizecolorrampsshowinghighdensityareasinredand
lowdensityareasinblue.Whileheatmapscanprovideaspatialsummaryofthedata,theyfall
shortoftemporalanalysis.Thisisparticularlyevidentwhenpointdataisstacked(samples
occurringatthesamelocationthroughtime)anditisnecessarytorepresentdatatrendswith
respecttotime.
Animationsprovideapowerfulmethodforportrayingpatternsorprogressivechangeover
time;suchasthoseexhibitedbychangesinsampledistribution,perspective,dataattribute
levelorgeography.Separatedatasetsmaybegeneratedforperiodsoftimewithinadataset
(daily,weekly,yearly,etc.),andcanthenbeanalyzedseparatelywithresultspresentedasa
seriesofmapscapturedinananimation.Whilethisisaneffectivemethodforvisualizing
temporalandspatialdatatrends,howthedataisdividedissomewhatarbitraryandmultiple
animationsmaybeneededtouncovertemporalandspatialpatterns.Additionally,thefocusof
animationscanbeplacedtooheavilyonthebeginningandendstagesofthedatawithreduced
emphasisonintermediatesteps.Audiencesarethereforerequiredtoremembermultipledata
iterationsatagivenpointtounderstandassociatedtrends.Asaresult,managementdecisions
basedonanimationsmaybeweightedtooheavilyonthebeforeandaftercomponentsofa
studyversusgeneratingpoliciesthatreflectanunderstandingoftrends.

14

Thegoaloftimeseriesanalysisistopresentdatawithsignificantspatialandtemporalrangesin
aneasilyunderstoodformatfordiverseaudiences.Onetoolfortimeseriesanalysisisthe
ArcGISHotSpotAnalysistool.ThetoolusestheGetisOrdGi*statistic,andcanbesetto
performanalysiswithinaspacetimewindowthatisolatesneighboringsamplingeventsand
illustrateshotandcoldspotsatintervalswithintheprojecttimeline.Theresultingoutputlayer
indicatesareaswhereobservedspatialclustersofhighorlowvaluesaremorepronouncedthan
expectedfromarandomdistributionofthosevalues(Figure3).

Figure3:OutputlayerfromHotspotAnalysistool;statisticallysignificantclustersofhighvalues(hotspots)and
lowvalues(coldspots)areindicatedbylargerredandbluecirclesrespectively

Hotspotanalysesareenhancedbycombiningthetoolslayeroutputwiththreedimensional
(3D)renderingofthedataonasurfacelikeArcGlobe.Tomaximizetheeffectivenessofthis
visualizationtechnique,itisadvisabletofirstrunthehotspottoolandthenaddafield
representingthetimeintervalbetweensamples.Specifically,thistimelapsefieldrepresents
timepassedbetweenthefirstsamplingeventandeachfollowingevent.Thisfieldisusedasa
multiplierinamathematicalexpressiontoextrudepointfeaturesona3Dplainandreflect
temporalprogression(Figure4).In3D,datatrendscanbemoreeasilyvisualizedbecausethe
addedelevationallowstheaudiencetoseparatethesamplingeventsofonetimeperiodfrom
another.Byextrudingfeaturesbasedonatimelapsefield,itbecomesclearerwhichfeatures
arerelatedandwhichareseparatedbytime.3Dvisualizationismosteffectivewithasmaller
studyareawhenyouhavealimitednumberoffeatures.


Figure4:ExampleofHotspotAnalysisoutputcombinedwith3DdatarenderinginArcGlobe

Aswithanypresentation,thepreferredmethodofdatavisualizationmustbepairedwiththe
intendedaudience,anditisoftenacombinationoftechniquesthatmustbeutilizedto
effectivelyengagetheaudienceandleaveanimpressionofdatatrends.Thesearejustafew
toolsthatmayhelpthevisualizationoftimeseriesdata.


Visualizing highly spatial and temporal data
KimberlyHyde,NationalMarineFisheriesService,Narragasett
Satelliteremotesensinggeneratesanextraordinaryamountofhighlyspatialandtemporaldata
thatcanprovideviewsoftheglobaloceannotattainablebyshipbasedsampling.Visualizing
qualitativeandquantitativedetailsfromtheselargedatasetsandcomparingthedatawithin
situormodeleddatacanbeachallenge.Seasonallyaveragedcompositesandanimationsare
twowaystoqualitativelycompareandcontrastthetemporalandspatialvariabilityofmultiple
satelliteproducts.Toachieveamorequantitativeanalysis,acommonpracticeistoextract
regionalmeansfrommappedsatelliteimagery.Simpletimeseriesplotscanshowregional
differencesamongthesubareasinagiventimeperiod,howevercomparingseveralregions
overmultipletimeperiodscanbecomedifficult(figure1a).Analternativeapproachistousea
colorscaletorepresentthemagnitudeoftheproductandtostackthetimeseriesinawaythat
onecaneasilycomparetheintraandinterannualvariability(figure1b).Theexamplefigure
providedshowsseasonalchlorophyllavariabilitywithinaregion(e.g.thereisageneralpattern
ofincreasedchlorophyllconcentrationduringthespringmonthsintheScotianShelfregion),
interannualvariabilitywithinaregion(e.g.differencesinthetimingandmagnitudeofthe
springandfallpeaksofchlorophyllintheGulfofMaine),andregionaldifferences(e.g.the
differencebetweenthesummerchlorophyllconcentrationsintheGeorgesBankandMiddle
AtlanticBightsubareas).
2

Figure1:Twoexamplesofhowtocompareandcontrasttimeseriesdataextractedfromsatelliteimagery:a)
showsanannualtimeseriesasaseriesoflineplots,andb)usesacolorscaletorepresentthedatavalueand
stacksmultipleyearsfromagivenregiontogether.

Asecondwaytocomparespatialdatasets(i.e.satellitedatatomodeloutput,datafromtwo
differentsatellitesensors,satellitedatafromtwodifferenttimeperiods)istocalculatea
differenceorratioanomaly(Figure2).Theratioanomalyisthepreferredmethodforlog
normallydistributeddatasuchaschlorophylla.Intheexamplebelow,theremotelysensed
chlorophyllconcentration(Figure2a)iscomparedtooutputfromabiogeochemicalmodel
(Figure2b).Theratioproduct(Figure2c)indicatesthatthemodeloutputhashigher
concentrations(theyellowtoredcolors)comparedtothesatellitedatathroughoutmostofthe
modeldomain.Thetwodimensionalhistogram(Figure2d)comparesthesatelliteandmodel
outputdatainaxyplotandusesacolorscaletorepresentthehistogramfrequency.

Figure2:Compositeofa)satellitedata,b)modeloutput,c)theratioanomalybetweenthesatellitedataand
modeloutputandd)atwodimensionalhistogramcomparingthesatellitedataandmodeloutput.Thisfigure
waspreparedbyK.HydeandadaptedfromHofmannetal.(2011,AnnualReviewofMarineScience3(1):93
122).

Amajorchallengeofhighlyspatialandtemporaldataiscreatingqualitativeandquantitative
visualizationsthatsummarizethedatawhilealsoretainingthefinescaleresolution.Thetwo
examplespresentedhereworkwellinastaticsetting(e.g.figuresinamanuscript),however
whenthistypeofdataispresentedinotheroutletssuchaswebsitesandkiosks,developers
shouldworktocreatevisualizationtoolsthatallowtheusertointeractivelyexplorethismulti
dimensionaldata.


Challenges, lessons learned, and a FEW potential visualization solutions for
synthesizing science to inform Management in south Florida
ChristopherKelble,AtlanticOceanographic&MeteorologicalLaboratory
ThetransitionfromsinglesectororsinglespeciesmanagementtomultisectoralEcosystem
BasedManagement(EBM)requiresthatweprovidemanagementwithusefulvisualizations
thatsynthesizeourintegratedscientificknowledgeaboutthebiophysicalandhuman
dimensionsoftheecosystem.Thefirststepinthisprocessistodeveloptoolsthatintegrateour
knowledgeacrossdisciplinesinamannerthatwillbeusefultoEBM.Tothisendwehave
developedDriver,Pressure,State,EcosystemService,Response(EBMDPSER)conceptual
modelsthatsynthesizeourbiophysicalandhumandimensionsknowledgebyhighlightingkey
ecosystemcharacteristicsandtheirconnectionstoeachotherandhumansociety.Thereare
often20ormorecomponentsineachmoduleresultingingreaterthan100ecosystem
4

componentsintheconceptualmodelandgreaterthan1,000connectionsamongjustthe
pressures,states,andecosystemservices.

FigureCK1.TheinfographicforBarrierIslandecosystemsinsouthwestFloridashowsthedominantpressures
uponthekeystatecomponentsandhumanuseswithintheecosystem.

Giventheimmensenumberofcomponentsandconnections,itisdifficulttovisuallyrepresent
thesemodelsinamannerthatiseasilyunderstood,butstillconveysimportant,useful
information.Onewaytoattempttoachievethisgoalisthroughinfographics(Fig.CK1).These
infographsicshowthroughcartoonimagesthemajorpressuresimpactingtheecosystem,the
keycomponentsofthestatemoduleandhumanusesintendedtoyieldthebenefitsfrom
ecosystemservices.Wediscussedhowtheseinfographicscouldbemodifiedormadeinto
moviestodepicttheaffectsofpressuresandpotentialmanagementscenarios.
ThemaingoalofEBMistosustainablyproducetheecosystemservicesthatbenefitsociety.To
accomplishthisgoalwemustquantifythecumulativeeffectofmultiplepressuresuponthese
ecosystemservices.Quantifyingtheconnectionsbetweenpressures,states,andecosystem
servicesintheEBMDPSERmodelleadstothedevelopmentofanetworkmodels.Thisnetwork

FigureCK2.Abarchartdepictstherelativeimpactofeachpressureonallecosystemservices.

modelcanbeanalyzedtoconductaholisticriskassessmentthatdeterminesthecumulative
effectofallpressuresonecosystemservices.Itcanalsobeusedtodeterminethepressures
causingthemoststressontheentirerangeofecosystemservices.Thisinformationsummarizes
theimpactofamultitudeofpressuresuponamultitudeofecosystemservicesmakingit
difficulttovisuallydisplay.Oneoptionsistodevelopabarchartdepictingthecumulative
impactofeachpressureuponallecosystemservices(Fig.CK2).Theinversecanbedoneto
showthedegreetowhicheachecosystemserviceisimpactedbythemultitudeofpressures.
However,theseareoflittleutilityifyouaremakingadecisionbaseduponasingleecosystem
serviceorpressure.Inthiscaseaspidergraphcouldbeusefulwhichshowstherelativeimpact
ofeachpressureoneachEcosystemService(Fig.CK3).Theseradarplotsquicklybecome
difficulttointerpretasthenumberofvariableincreasesandanewvisualizationtechnique
wouldbepreferred.Thesevisualizationexamplesarepreliminaryattemptstoconveyscientific
synthesistomanagersattemptingtoimplementEBM.Itisintendedthatalloftheseproducts
willberefinedandimprovedthroughinteractionswithmanagersandasvisualization
techniquesadvance.

FigureCK3.Aradarplotshowstherelativeimpactofeachpressureoneachecosystemservice.


Dynamic conceptualization of habitat
JohnManderson,NEFSC;GavinFay,NEFSC;EricFredj,JerusalemCollegeofTechnology;Josh
Kohut,RutgersUniversity;ScottLarge,NEFSC
Habitatsintheseaarefundamentallydefinedbypropertiesandprocessesoftheoceansfluid.
Carefullydesignedecologicalmodelsincludingspeciesnichemodelscannowbeintegrated
withhydrodynamicmodelstodevelophydrodynamicinformationsystems(HIS).HIScanthen
6

beusedtovisualizeandanalyzethedynamicsofkeyecosystemprocesses,includingseascape
andhabitatdynamicsinthewaythatgeographicinformationsystems(GIS)areusedtovisualize
andanalyzeterrestriallandscapeandhabitatdynamics.WehavedevelopedaHISthat
integratedasimplethermalnichemodelforanimportantforagespeciesinthemidAtlantic
Bightwithtemperaturehindcastsfrom19572007derivedfromaRegionalOceanModeling
Systemmodel.TheHISallowedustovisualizethermalhabitatdynamicsandtocalculate
statisticsdescribingtemporaldynamicsofspatialhabitatcharacteristicsthatcouldinfluence
importantprocessesregulatingthedynamicsoftheregionalpopulationaswellasstrongly
interactingspecies.Weperformedpreliminarystatisticalanalysestoinvestigaterelationships
betweenseascapeindicesandecosystemresponseswithinaDrivingforcePressureState
ImpactResponse(DPSIR)indicatorframework.


Designing geospatial data visualizations for a general audience
TomButkiewicz,CenterforCoastal&OceanMapping,UniversityofNewHampshire
Dynamicoceansimulationsaregeneratingincreasinglycomplexmultilayer3Doceanmodels
containingdepthvaryingflowvectors,temperature,salinity,etc.However,many
oceanographersandothermarinescienceresearchersarestillusingtraditional2Dsoftwareto
consumethisdata.Properlydesigned3Dvisualizationtoolscanbehighlyeffectivefor
performinganalysiswithin,andrevealingthecomplex,dynamicflowpatternsandstructures
presentinthesemodels.
CCOMsexperimentaldynamicoceanvisualizationsystemincorporatestheperceptualbenefits
ofstereoscopicrendering,tobestrevealandillustrate3Dstructuresandpatterns,andmulti
touchinteraction,allowingfornaturalandefficientnavigationandmanipulationwithin3D
environments.Exploratoryvisualanalysisisfacilitatedthroughtheuseofahighlyinteractive
toolsetwhichleveragesasmartparticlesystem.
Configuringparticleswithspecializedbehaviorscanprovidelimitedsimulationcapabilities.This
includespollutantreleasescenarios,suchaspredictingthepathoftheoilplumeinthe
DeepwaterHorizonoilspill,orthepathofradioactivecoolantfromtheFukushimanuclear
disaster.Habitatmappingcansimilarlybesupportedbymodelinglarvaltransport,etc.(Figure
1).
Otherusesincludeplanningsurveymissionswithinforecastedflowmodels,toplotcoursesthat
takeadvantageofcurrentstomaximizeenergybudgets.Realtimedatafromremotesensors,
suchastheARGOfloatnetwork,canbedirectlycomparedtothesimulationspredictions,
allowingformodelvalidationanddetectionofsensormalfunctions(Figure2).
7

Thissystemwasdesignedtobemoreintuitive,andallowfordirectinteractionwithonscreen
elements.Comparedtomostcommercialanalysispackages,itrequiresfarlesstrainingtouse,
makingiteasierforthegeneralpublictoutilize.Theinterfaceencouragesexploratoryanalysis,
wheretheusercanquicklytryoutmanydifferentideasandanalyseswithouttheburdenoffirst
rigidlydefiningparameters(Figure3).Thiscanleadtothediscoveryofmoreinsightsthat
mightotherwisebemissed.
ThisstyleofintuitiveexploratoryanalysisisalsocentraltothesuccessoftheUrbanGrowth
DecisionSupportSystem.TheUGDSSallowsstakeholdersandothermembersofthegeneral
publictoperformsimplespatialanalysesontheoutputofacomplexurbangrowthsimulation.
Byprovidinganumberofeasytousetoolsforvisuallyqueryingthedata,theuserscan
investigatetheirownregionsofinterest(Figure4).Whentheusercomparesregionsof
interest,thesystemautomaticallydetectsanypotentiallymisleadingfactorsthattheuser
mightbeunawareof.Italertstheusertotheseissues,andprovidessemiautomatedmethods
forrefiningtheirregionstorectifythem(Figure5).

Figure1.Interactivepathediting.


Figure2.RealtimedatafromRTOFSARGOfloatsfacilitatescomparison.

Figure3.Multipolepathcomparisons.


Figure4.Ecosystemusecomparisons.

10


Figure5.Exampleworkspaceofregionsofinterestcomparisons.

11

Dealing with scientific and communication uncertainty through data


visualization
HowardTownsend,NMFS/OHC/NCBO
Ecosystembasedmanagement(ebm)iscomplexandfraughtwithuncertainties.Ecosystems
arecomplexandinteractionsarenonintuitive.Keepingtrackofthemwithdataismessy.Trying
tounderstandandmanagethemisdifficultbutnecessary.Understandingtheuncertainty
associatedwithebmandproperlycommunicatingtheuncertaintyshouldhelpthetransition
towardsebm.Ecosystemscientistsandmodelershavetakenstepstocategorizetheuncertainty
associatedwithebm(Figure1).
Modelershavedevelopedapproachesforvisualizingdataandmodeloutputtodemonstrate
thescientificuncertainty(e.g.,naturalvariability,observationerror,andstructuralcomplexity);
however,somemanyoftheseapproachesresultinfairlycomplexoutput.Typicalapproaches
foruncertaintyinclude:usingMonteCarlosimulationstoincorporatenaturalvariabilityand
observationerror,usingmultiplemodelstoaccountforstructuraluncertaintyinmodels,and
usingmultiplescenariostoaccountforoutcomeuncertainty.Theseapproachesareveryuseful
forscientificuncertainty,butoftenaddtothecomplexityofoutputpresentedtomanagersand
mayexacerbatemanagementassociateduncertainty.
Ecosystemscientistsandmodelersalsoneedtoconsiderthemanagementuncertainties(e.g.,
specificityandcommunicationuncertainty).Mostmanagersarefamiliarwithsinglesectors
(fisheries,waterquality,etc.),butasebmisanewparadigm,sometimesmanagersmental
models,basedonsinglesectormanagement,conflictwithwhattheecosystemmodelsare
showing.Overcomingthiscognitivegaprequireslotsofcommunicationsanddifferent
approachesofdemonstratingdataiterativetwowayinteractions(theSubwaymethod)and
sometoolsthatallowdataexploration.Thisisdifficultbecauseoftimeconstraintson
managersandmodelers.
Interactivedatavisualizationtechniquesholdpromiseforfacilitatingiterativetwowaydata
andmodeloutputexplorationandovercomemanagementuncertainties.Frequentlywithstatic
presentations,ecosystemscientistsandmodelersarelimitedtodemonstratingportionsofa
modeloutputtocommunicatefocalpointsforebmtechnicalguidanceforagiven;however,
staticrepresentationofcomplexecosystemsdoesnotallowtheflexibilityofdemonstrating
otheraspectsoftheecosystem.Asafocalpointofanecosystemisdemonstratedtomanagers,
questionsaboutotherportionsoftheecosystemmayarise.Developingstaticvisualizationsof
allcomponentsofcomplexecosystemsisnotfeasible,sousingdynamicvisualizationtools
allowsflexibilitytorespondtomanagersquestionsandfacilitatesaniterative,twoway
explorationofdataandmodeloutputs.HansRoslingprovidesagooddemonstrationof
12

interactivedatavisualizationanddataexploration(http://ed.ted.com/lessons/hansrosling
showsthebeststatsyouveeverseen).

Structural
complexity
Natural
variability

Observation
error
Ecosystem model

Natural aquatic
system

What
we
know

Sampling, data collection

What we
dont
know

Communication
uncertainty
Resource use

Management
objectives
Stakeholders
Outcome
uncertainty

Unclear
objectives

Regulations
(closed areas,
harvest rates, )

Ecosystem managers

Figure1.Sourcesofuncertaintyassociatedwithebm.


NEFSC Data Server: access & visualization with Python
JimManning,NOAA/NEFSC
Thesimpleideaofpostingcomputercodeinadditiontodatawasexpressed.Whileweare
promotingtheuseofPYTHONcodeasanalternativetoMATLAB,theactuallanguageis
secondarytotheprimaryideaofsharingcode.Thetalkoutlinedthehistoryofoureffortsfrom
theinitialformationoftheWoodsHolePythonUsersGroupinmid2011throughourpresent
daydirectionsonhowwehopetoproceed.Thiswasanintroductorypresentationleadingup
torelatedandmoredetailedinformationtobedeliveredinSignell'sandDiStephano's
subsequenttalks.
Asapointofmotivation,itisdifficulttobuildawebsitethatcanbeusefultoalltypesofusers.
GiventhequantityofthedatathatisservedthroughtheNEFSCOceanographyBranchwebsite
(profiles,trawldata,mooringtimeseries,driftertracks,planktondistributions),forexample,it
13

ishardtoprovideagraphicaluserinterfaceandmappingtoolthatwouldbeidealforall
disciplinesandlevelofscientificinquiry.Instead,wehopetopointtorepositoriesofopen
sourcecodethatuserscandownloadandrunontheirownmachine.Thiscodewillallowthem
tobothaccessthedataaccordingtospecificcriteriaandplotitinavarietyofways.Withalittle
experimentationandpractice,theyshouldbeabletomodifythecodetogetexactlywhatthey
need.
AnothermotivationfortheprojectarisesfromtherecentcallforSTEMactivitiesinthe
classroom.Sincewearenowfocusingonopensourcesoftware,wecanoffertheseroutinesto
teacherswhocaninstructtheirstudentstodomorethannavigatecomplicatedwebsites.They

Figure1.Observedandmodeledwatertemperaturebasedinlobstertrapsensor.

canactualinstalltheprogramsontheirmachines,learntorunsimpleplottingpackages,and
perhapsbegintoappreciatetheartofcomputerprogramming.Weareintheprocessof
writingaproposaltoNOAA'sEnvironmentalLiteracyGrantProgramtosuggestthisstrategyof
exploringNOAAdata.
Whileitwasnotpossibletorunattheconference,acoupleofsimpleexampleswereprepared
todemonstratetheidea.Oneprogramaccessesbothobservedandmodeledoceanbottom
temperaturesfromaparticularsiteandplotstheirtimeseries(figure1).Anotheroverlays
multiplevariableslikeSST,CODARsurfacecurrents,anddriftertracksonamap.Inallcases,the
dataisstoredonremotemachinesandisaccessibleviaIOOSdatastandards.Whilestillunder
developmentandnotwelldocumented,thecodeisstoredonthejamespatrickmanning
GitHubcoderepositorywithSphinxdocumentation.

14

Web access, analysis and visualization of standardized oceanographic and


meteorological data
RichSignell,USGS
Thistalkfocusedontheimportanceofwebservicesandcommondatamodelstoallow
developerstoefficientlyandeffectivelywriteapplicationstoaccessstructuredand
unstructuredgridmodeloutputandotherstandarddatatypes,suchastimeseries,profiles,
swathsandtrajectories.AsystembasedonUnidatatechnologieshasbeendeployed
throughouttheIntegratedOceanObservingSystem(IOOS)andhasmadedozensofmodels
throughoutall11IOOSregionsavailablethroughcommonservices,withoutforcingmodel
providerstorewritetheirnonstandardmodeloutput(Figure1).AlayerofXML(NetCDF
MarkupLanguageNcML)iscreatedthatmapsexistingdataintothecommondatamodelby
providingmissingmetadata,whichthenbecomesavailableviaavarietyofstandardservices.
ThesuccessofthistechniquewithmodeloutputhasledIOOStostartdeliveringtimeseries,
profilesandotherdataviathistechnique.ThisallowsgenericapplicationslikeIDVtoaccess
datafrommanydifferentmodelswithoutregardtomodelspecifics(Figure2).

Figure1.ModeldatainteroperabilitydiagramforIOOS.

15

Figure2.VisualizationoftwocommondatamodeldatasetsstreamlinesfromtheWRFatmosphericmodeland
bottomtemperaturesandasalinityisosurfacefromtheUniversityofMainePOMoceanmodel.

16

Oceanographic data and dataservices at PacIOOS, a regional association of


IOOS
JamesPotemra,PacIOOS
TheIntegratedOceanObservingSystem(IOOS)isorganizedaround11regionalassociations
(RA).ThePacificIslandsOceanObservingSystem(PacIOOS)istheRAfortheinsularPacific,
includingtheStateofHawaii.ThemainpurposeofPacIOOSistoprovidetimely,useful
informationabouttheoceanenvironmenttoawidevarietyofusers.PacIOOSisconsideredan
endtoendsysteminthesensethatwiththesystemareoceanobservingplatforms
(autonomousgliders,buoys,moorings,etc.),numericalmodels,andremoteobservations,but
PacIOOSalsomaintainsasetofdataserversthatprovidethedata(anddataderivedproducts)
tothegeneralpublic.
Theforecastmodelsandobservationsaregearedaroundprovidinginformationinfourgeneral
areas:coastalhazards,maritimesafety,waterqualityandoceancirculation.Usersofthis
informationincludeeverythingfromrecreationalswimmers,fishermen,managementand
planningentities,searchandrescuepersonnel,etc.
WebaccessisobviouslyakeycomponenttothePacIOOSdatasystem.Themainaccess
mechanismisviaamapbasedbrowsingtool.ThebackendsystemisbuiltuponOPeNDAP
services,mainlyTHREDDS.THREDDS,orthematicrealtimeenvironmentaldistributeddata
system,wasdevelopedatUnidataandprovidesdirect,binaryaccesstodata.AtPacIOOS,
THREDDSisusedtobothtoexposeallthedatabutalsotoprovidedataviaservicestothe
PacIOOSviewers.
Themainviewer,http://pacioos.org/voyager,allowsuserstogenerateoverlaysofany/all
PacIOOSdatainamapbasedview(seeFigure1).Thedataarelistedattheleftofthescreen,
anduserscanselectspecifictimes,geographicand/ortemporalrangesandgenerateacustom
map.ThebackgroundisbasedontheGooglemapAPI,andmostoverlaysaregeneratedonthe
flybytakingadvantageoftheTHREDDSwebmapservice(WMS).Onceuserscreateamapof
interest,thereareoptionsatthetoptoeithersavetheimageforprintingorsavingalinktothe
exactimage(e.g.,toincludeitintoawebpageortoemailtosomeone).
Inadditiontothisdatabrowsingtool,PacIOOSalsooffersageospatialserverbasedon
GeoServerandOpenLayers(http://pacioos.org/geoexplorer)andseveralpregenerated
informationanddatabasedproducts,suchaparticletrajectorymaps,highwaterforecasts,
inundationplots,andmore.

17


Figure1.ExamplemapoverlayfromPacIOOSVoyager(http://pacioos.org/voyager).

18

Data and visualization integration via web based resources


MassimoDiStefano,WHOI
Developmentofcyberinfrastructuretofacilitatecollaborationandknowledgesharingfor
marineIntegratedEcosystemAssessments(IEAs).
Themaintoolisbasedonawebapplication(IPythonNotebook)thatprovidestheabilityto
workonverydiverseandheterogeneousdataandinformationsources,providinganeffective
waytosharethesourcecodeusedtogeneratedataproductsandassociatedmetadataaswell
assaveandtracktheworkflowprovenancetoallowthereproducibilityofadataproduct.

Here,isasimplifiedschematicofthesoftwaredesign,startingwithasourcedatasetandending
withafinalproductforanEcosystemStatusReport.
Akeyfeatureisthatmetadata,embeddedinthefinalproduct,areacquiredduringthe
processingandplottingofthedata.
Inthiswayweareabletorecordtheprovenanceneededtoreproducethedataproducts.
WeareusingtheIPythonNotebookastoolforcollaborativedataProcessing,workflow
ProvenanceandprodutsPublishing.
IPython(InteractivePython)canberuninteractivelyoverthewebprovidingtotheuseran
effectivewaytoworkonhisdata.
HereinanexamplesessionshowingtheIPythonNotebookinterfaceusedtoruninteractively
thecodetoproducesomefiguresfortheNESLMEEcosystemStatusReportandtoexecute
somegeospatialdataanalisysusingtoolslikeGRASSGISasGeographicInformationSystem,R
forstatisticalanalysisincombinationwithotherfreeandopensourcesoftwaretools.

19

20

Visualizing Fitness for Purpose


RobertGroman,WHOI
TheBiologicalandChemicalOceanographyDataManagementOffice(BCODMO)worksin
partnershipwithoceanscienceinvestigatorstopublishdatafromresearchprojectsfundedby
theBiologicalandChemicalOceanographySectionsandtheOfficeofPolarProgramsAntarctic
Organisms&EcosystemsProgram(OPPANT)attheU.S.NationalScienceFoundation.Since
2006,researchershavebeencontributingdatatotheBCODMOdatasystem,andithas
developedintoarichrepositoryofdatafromocean,coastalandGreatLakesresearch
programs.TheendgoalsoftheBCODMOaretoensurepreservationofNSFfundedproject
dataandtoprovideopenaccesstothosedata.
BCODMOhasdevelopedanendtoenddatastewardshipprocessthatincludesallphasesof
thedatalifecycle:(1)workingwithinvestigatorsattheproposalstagetohelpthemwritetheir
datamanagementplan;(2)registeringtheirfundedprojectatBCODMO;(3)addingdataand
supportingdocumentationtotheBCODMOdatarepository;(4)providinggeospatialandtext
baseddataaccesssystemsthatsupportdatadiscovery,access,display,assessment,integration,
andexportofdataresources;(5)exploringmechanismsforexchangeofdatawith
complementaryrepositories;(6)publicationofdatasetstoprovidepublishersofthepeer
reviewedliteraturewithcitablereferences(DigitalObjectIdentifiers)andtoencourageproper
citationandattributionofdatasetsinthefutureand(7)submissionoffinaldatasetsfor
preservationintheappropriatelongtermdataarchive.
Weprovidebothatextbasedandgeospatialinterfacetothedatawemanage(see
http://www.bcodmo.org).Ouruseofvisualizationtoolsandtechniquesaretheretohelp
potentialusersofthedatawemanagetofindthedata,assessfitnessforpurpose,and
eventuallydownloadthedatatheyneedfromoursite.Tothisend,weofferseveraltypesof
plotsandvisualizations,consistentwiththetypeofdatabeingviewed,includingXYplots
(Figure1),biologicalabundanceplots(Figure2),timeseriesplots(Figure3),andexportedKML
filestoGoogleEarth(Figure4).Contacts:RobertC.GromanandM.DickyAllison
Alistofourlessonslearnedandrecommendationsincludethefollowing:
Agooduserinterfacetodatashouldhaveeffectivevisualizationtoolstoaidtheinvestigatorin
determiningwhetherdatawillbeusefultothem.
CanwebeAllthingstoallpeople?Wecantry,andoneoftheimportantrequirementsis
informationaboutthedata.
Toolsmustbedatadriven,andmetadataenablesthis.

21

Attributesofdatafitnessincludedatalocation,datatype,acquisitiondateandtime,dataand
metadataquality,andthemethodologyusedtocollectandprocessthedata.
Afundamentalprerequisiteforonlinedatavisualizationisthatinvestigatorsmustbewillingto
sharetheirdata.
Therichertheonlineoptions,themorecomplexthesystemistouse.Helptext,videohelpfiles,
andcarefulprogramming(suchasgooddefaultvalues)canmediatethis.
Recognizethatdifferentaudiencesrequiredifferentinterfaces.Thisaddstodevelopmenttime
andcost,andcanreduceonesfocus.

Figure1:XYplotFigure2:Abundanceplot

Figure3:TimeseriesplotFigure4:ExportedKMLfiletoGoogleEarth

22

Visualizations of Habitat and Siting Decisions


MichaelParke,NationalMarineFisheriesService
ThePacificIslandsFisheriesScienceCenterconductsavarietyofvisualizationeffortsthatare
designedtosupporthabitatdelineationandanalysis,stockassessments,ecosystembased
managementandmarinespatialplanning.Ourbenthichabitatmappinggroupusesmultibeam
sonars,multiplecameraplatformsandinstruments,anddiverobservationstogeneratedata
thatarevisualizedandanalyzedusinggeographicinformationsystems.Atypicalproductfroma
multibeamdataanalysiswouldbearepresentationofgeomorphologicalstructuresusinga
secondorderbathymetryderivativecalledbathymetricpositionindex.Byexplicitly
incorporatingdataextractedfromtowedvideos,wewereabletovisualizebiologicallysensitive
areasthatshouldbeavoidedduringanchoringinSaipan.

23

PIFSCmodelersandvisualizationexpertsstrivetodeliverproductsthatuseappropriate
temporalandspatialscales,containadequatedocumentationandmetadata,andoffersome
measureofuncertainty.Ongoingeffortstovisualizemoreephemeralphenomenathatare
neededtoimprovestockassessmentsandecosystembasedmanagementsuchasocean
currentsandtemperatures,boundarylayers,larvaltransport,andlifehistory/habitat
associationsareconstrainedbylackofdata,especiallyatfinerscales,andmodellimitations.
OneotherongoingvisualizationeffortisthePIFSCAquacultureMarineMapper
(www.pifsc.noaa.gov/MarineMapper),awebbasedscreeningandvisualcommunicationtool
designedtosupportresponsibleoffshoreaquaculturedevelopmentintheMainHawaiian
Islands.Themainpremiseofthetoolisthatinteractivemapscanbeusedtocommunicate
oceanographic/ecologicalandocean/landusepatternsandthusempowerpeopletoexplore
theirworld,tocommunicatetheirvaluesandvisionforthefuture,andtoengagewithothers
aboutthesevaluesandvisions.
Thetoolprovidesaccesstopubliclyavailabledata(andmetadata),allowsuserstoviewany
combinationofthesedatathattheyneed,developcustomizedmaps(visualizations),andruna
firstpasssiteselectionmodelusinguserdefinedparameters.Itallowsuserstoidentifywhich
areashavepotentialforaquaculturebasedontheparametersthatareimportanttothem.The
siteselectionmodelisprimarilydesignedtofacilitatecommunicationacrossusergroups
(entrepreneurs,regulators,communitygroups)bymakingvaluejudgmentsexplicitand
providingimmediatevisualizationoftheimpactsofsuchjudgments.Amajorlimitationofthe
toolisitsinabilitytovisualizeorcalculateriskorvulnerabilityofnaturalsystemstoany
particularaquacultureoperation.

24

Habitat modeling and visualization


ScottGallager,WoodsHoleOceanographicInstitutionandtheHabCamGroup
Habitatcharacterizationandpredictivemodelingrequiresintegrationofinformationatmultiple
spatialandtemporalscalesonbenthicandpelagicbiologicaldistributions,substrate
composition,geomorphology,watercolumnhydrography,aswellasmanyotherenvironmental
featureswhichprovidesuitableconditionsfororganismstoflourish.HabCamisatoolthat
providesauniqueglimpseoftheseafloorthroughopticalandacousticimaging.TheHabCam
vehiclefliesovertheoceanbottomtakingsiximagespersecondcreatingnoninvasive
continuousimageribbonsandenvironmentaldatainrealtime.Habitatmodeling(Fig.1)begins

Figure1.Flowchartfordevelopinghabitatsuitabilitymapsforsinglespeciesandbiologicalcommunities.

withcreatingstandardsizegeolocatedgridsofpredictorvariablessuchasbathymetry,
geomorphology(slope,rugosity,andgradientfrombothopticandacousticdata),substrate,
epifaunacover(e.g.,lacytubeworm,bryozoa,encrusting,sponge,tunicate,etc.)temperature,
salinity,etc.Thedistributionofpointpresencedataforbiologicaltargetssuchasgroundfish,
scallops,seastars,etc.isthenoverlaidonthepredictorgrids.Statisticalmodelssuchasgeneral
additive,linear,randomforests,ormaximumentropyarethenusedtodevelophabitat
classificationsbasedontrainingdata.Forthisstudywedevelopedavariationofthemaximum
entropymodelMaxent(Phillipsetal.,2006),wherebybothpresenceandabsencedata
contributedtothepredictedmodelresults(Fig.2).Themodeloutputforeachtargetspecies
wasthenspatiallyclusteredintofourormorebiologicalcommunitiesusingallofthepredictive
rasterandpresencedata.Accuracyofthemodelresultsarecrossvalidatedusingleaveoneout
procedures.Theresultsaremapsofpredicteddistributionsofindividualspeciesand
25

communitiesthatmaybeviewedovertimetoassessvariationasafunctionofenvironmental
change.

Figure2.PredictedsuitablehabitatforadultcodandhaddockinthenorthernsectionofClosedAreaII,Georges
Bankandregionstothewestoftheclosureboundary(blackline).Noteabsenceofsuitablehabitatalongthree
sandridgesextendingnorthwesttosoutheast.

26

Communicating an understanding of lower trophic levels in marine


ecosystems
AndyPershing,UniversityofMaine
ThenorthwestAtlantic,includingtheNEShelfLMEandtheCanadianShelfandLabradorSea
arechangingrapidly.Inparticular,theLabradorSeaiswarmingfasterthananyregioninthe
globalocean.Thechangesinthephysicalsettinghavefarreachingimpactsonecosystems
throughouttheregion,affectingprimaryproductivity,planktoncommunitystructure,fish
populations,andfisheries.Visualizationisanimportanttoolforunderstandingand
communicatingthecausesandconsequencesofthesechanges.Formanyglobalchange
problems,thekeychallengeistohighlightcomponentsthatarechangingrapidly.Removing
annualcyclesexposeslargedeviationsfromexpectedconditions.Whenvisualizingthese
deviations,usingacolorschemethatemphasizesthesignratherthanthemagnitudeisuseful.
Inthislight,the2012SSTanomalystandsoutasthemajorfeatureintheNorthAtlanticlast
year,andoneofthelargesteventsever.Whencommunicatingwithbroaderaudiences,careful
thoughtisrequiredtodevelopvisualizationsthatquicklyconveytheintendedmessage.
However,ascientistexploringdataoftenwantstoseemanyvariablesplottedtogether.These
uglygraphics,forexamplealargetableofcorrelationsbetweenplanktonabundanceand
physicalindicators,canbeveryusefulforanindividualscientistforidentifyingpatterns,but
wouldbeoffputtingtomanyotheraudiences.Oncearesultisidentified,communicatingthat
result,regardlessofaudience,involvestellingastory.Visualizations,forexample,afoodweb
highlightingcorrelationsamongspecies,areanimportanttoolintellingthestory,buttheyare
notthewholestory.Anarrative,eitherthroughtext,video,oralivepresentation,areessential
forguidingtheviewer/readerthroughthestory.Animationsordiagramscanbeespecially
helpfulindevelopingthenarrative.

27

Find And Visualize NOAA Data: An Overview of the NOAA Earth Information
System (NEIS) and TerraViz
JeffSmith,EarthSystemsResearchLaboratory
TheNOAAEarthInformationServices(NEIS)isaframeworkoflayeredservicesdesignedtohelp
thediscovery,access,understanding,andvisualizationofdatafromthepast,present,and
future.ItincludesavisualizationcomponentnamedTerraVizthatisamultiplatformtool,
runningondesktops,webbrowsers,andmobiledevices.Thegoalistoingest"bigdata"and
convertthatinformationintoefficientformatsforrealtimevisualization.Designedforaworld
whereeverythingisinmotion,NEISandTerraVizallowfluiddataintegrationandinteraction
across4Dtimeandspace,providingatoolforeverythingNOAAdoesandthepeopleNOAA
affects.
TerraVizisbuiltusingtheUnitygameengine.Whileagameenginemayseemastrangechoice
fordatavisualizations,ourphilosophyistotakeadvantageofexistingtechnologywhenever
possible.Videogamesareamultibilliondollarindustry,andarequitesimplythemost
powerfultoolsforpushingmillionsofpointsofdatatotheuserinrealtime.Ourpresentation
illustrateddisplayingenvironmentaldatainTerraVizataglobalscale,visualizingregionaldata
inscenessuchasthefloodingoftheWashingtonDCareaorrotatingacoastalecosystemin
threeaxes,anddevelopingenvironmentalsimulations/gameslikeexploringtheoceanfloorina
submarine.

TheNEISbackendsimilarlytakeslessonsfromprivateindustryasweuseApacheSolrtoallow
facetedsearchofNOAAdata,muchassiteslikeAmazonandNetflixdo.

28

Webelievethattohaveanimpactonsociety,datashouldbeeasytofind,access,visualize,and
understand.Pleasecontactusifyouwanttoexploreincludingyourenvironmentaldatawithin
NEIS/TerraVizorifyouwanttotalktousaboutdevelopingcustomsimulationsorgamesto
showcaseyourimportantdata.
DetailedContact:NOAA/EarthSystemResearchLab/GlobalSystemsDivision,Boulder,
Colorado
[email protected];[email protected];Jeff
[email protected]
Webresources:NEIS/TerraViz:http://www.esrl.noaa.gov/neis/
Submarinedemo:http://www.youtube.com/watch?v=Ukaln8_ai3c


Getting to observing data usign the climatology toolkit
RileyYoungMorse,GulfofMaineResearchInstitute
Isthewaterwarmerthisyearthanitwaslastyear?Thisisoneofthemostcommonlyasked
questionsbyusersoftheNortheasternRegionalAssociationofCoastalandOceanObserving
Systems(NERACOOS).Tohelpanswerthisquestion,ateamofdataprovidersandproduct
developersatNERACOOSsetouttodevelopanonlineinteractiveclimatologyvisualizationthat
wouldenableusersofthesystemtoseethedataandanswerthatquestionforthemselves.
TheresultisatoolcalledtheNERACOOSWeatherandClimateDisplay
(http://www.neracoos.org/datatools/climatologies).Thetoolconsistsofaninteractiveline
chartthatdisplaysthetimeseriesclimatologydataasabackgroundshadedarea,withthemin
29

monthlyandmaxmonthlymeanscreatingtheupperandlowerboundaries.Themeanofthe
fullclimatologyisplottedaswelltomakeupthestaticbackground.Userscanchooseabuoy
location,parameteranddepth,andthebackgroundwillautomaticallyupdatewiththerelevant
climatologydata.Additionally,themeansofagivenyeararedisplayedabovethestatic
climatology,andcanbeeasilychangedusingnavigationbuttons.Bymousingoveranypointon
thechart,theprecisevaluesaredisplayedtotheuser.Additionally,thefulldatatablecanbe
displayedtoaccessanddownloadthedataforfurtheranalysisindesktoptoolsofchoice.

Figure1:2012dailyclimatologyofwatertemperatureat1mdepthforBuoyA

Thebuoydataprovidersdevelopedacommonprotocolforcalculatingthedaily,weeklyand
monthlyclimatologystatisticsforeachbuoy,parameterandeachdepth.Thestatistics
calculatedinclude:count,mean,median,mode,min,max,stddev,pct0.5,pct02.5,pct97.5,
pct99.5,IQR(25%,50%).ThefilesareoutputascsvfilesintoawebaccessibleFTPdirectory.
Additionalautomatedtasksincludecalculationofcurrentyeardailymeansattheendofeach
day,andmonthlymeansattheendofeachcalendarmonth.Therangeofavailabledatato
calculatetheclimatologieswasdifferentforeachbuoy,withsomegoingasfarbackas2001.
Thebackgroundclimatologieswillbeupdatedattheendofeachyear.
Onthewebinterfaceside,aJavaScriptplottinglibrarycalledFLOTwasusedtocreatetheweb
basedviewer.ScriptsweredevelopedtoparsethecsvfilesandoutputintoJSON(JavaScript
ObjectNotation)format,anobjectorienteddatainterchangeformatusedbytheJavaScript
FLOTlibrary.Thisprocessisautomatedforahighlyinteractive,uptodatedisplay.

30

Futureplansincludedevelopmentofamapbasedinterface,additionaldatasetsand
parameters,anddisplayofmorestatisticalinformationforadvancedusers.Thetoolhas
generatedtremendousinterestdemonstratingthevalueofdeliveringcomplexinformationina
simple,interactiveformattoreachandengageawidevarietyofusers.


Developing a climate data visualization tool for marine ecosystem managers
KarstenShein,NOAANationalClimaticDataCenter,Asheville,NC28801,
[email protected],(828)2714223
TheNationalOceanicandAtmosphericAdministrationisresponsiblebothforthecollectionand
archivalofavastproportionofthedatacollectedfromthenationsenvironmentalmonitoring
activities,aswellasprovidingstewardshipoverthenationsmarineecosystemsandthe
valuableecosystemservicestheyprovidetosociety.Inordertoensurethelatter,data
availableviatheformermustbeaccessibleandprovidedinwaysthatmarineecosystem
managerscanreadilycomprehend.
Overthepastfewdecades,theimpactsofclimatevariabilityandchangeonmarineecosystems
havebecomemoreacuteandareacknowledgedasasignificantthreattotheirsustainable
futures.However,althoughthereisagreatquantityofclimatedataheldbyNOAA,anda
varietyoftoolsbywhichthesedatamaybeobtainedanddisplayed,marineecosystem
managersarefacedwithseveralobstaclesintheirattemptstotransformrawclimatedatainto
meaningfulvisualizationscapableofsupportingsoundmanagementdecisions.
Chiefamongtheseobstaclesarethe(1)lackofuniformityindatasetaccessandformat,(2)the
requirementthatdatafirstbeobtainedbeforeitmaybevisualized,and(3)therequirement
thatmanagersnotonlyunderstandwhatdatatheyrequireandwheretheymaybefound,but
alsothattheyunderstandwhichofseveralsimilardatasetsmaybemostappropriateto
addressagivenmanagementquestion.Whatismore,thereexistfewdataaccesstoolsthat
includetheabilitytovisualizethedata.Thisisanimportantissue,becausethevisualizationof
dataoftenaidsauserindeterminingwhatdataaretrulyimportanttoaparticularmanagement
question.Withouttheabilitytovisualizethedata,amanagermustblindlydownloadall
availabledata,andthenexpungeanydatathattheirsubsequentvisualizationdeemsirrelevant.
Thispresentsanunnecessaryburdenonthemanagerstime,theserver,andonbandwidth.In
anefforttobetteraddresstheneedsofmarineecosystemmanagersandovercomethese
barriers,aNOAAwidegrouphascreatedtheIntegratedMarineProtectedAreaClimateTools
(IMPACT)project,withinwhichdataaccessandvisualizationtoolsarebeingdeveloped.The
overarchinggoalofIMPACTwastoensurethatdataaccess,visualization,andunderstanding
werebaseduponthemanagementquestionsbeingposed,andthereforecouldoptimizethe
31

resourcesneededtofacilitatetheidentification,access,visualization,anduseofappropriate
data.
WCTIMPACT
OnesuchtoolistheWCTIMPACTsoftware.NOAAsNationalClimaticDataCenter(NCDC)has
beenworkingwiththeGulfoftheFarallonesNationalMarineSanctuary(GOFNMS)Ocean
ClimateCentertoenhanceandexpandthefunctionalityofNOAAsWeatherandClimate
Toolkit(WCT).TheWCT(Fig.1)isafreelyavailable,Javabasedtool
(http://www.ncdc.noaa.gov/oa/wct/)designedtoaccessanddisplayavarietyofNCDCs
georeferencedclimatedataproducts(e.g.,satellite,radar,reanalysis).However,theWCT
requirestheusertohavealreadyidentifiedadatasetofinterestandgainedaccesstoit.This
canlimititsutilitybyuserswhoarenotknowledgeableaboutwhichdatasetsarerelevantto
theirneedsandwherethosedatasetscanbefound.TheIntegratedMarineProtectedArea
ClimateTools(IMPACT)prototypemodificationtotheWCTeliminatesthoserequirements.
Instead,WCTIMPACTcouplesauserqueryapproachwithaquasiexpertsystemthat
determines,retrieves,andloadstheappropriatedataproductsforvisualizationandanalysisby
theuser.Relevantdataproductsareidentifiedbasedontheenvironmentalvariablesinwhich
ecosystemmanagershaveindicatedareimportanttotheirecosystems.

Figure2.NOAA'sWeatherandClimateToolkitimageofNARRAnonconvectivecloudcover(%)overthePacific
CoastonJune17,2012at09:00GMT.

32

AtthecoreofWCTIMPACTisaneedsassessmentthatwasconductedthroughinteractiveand
iterativedialogbetweenNCDCprogrammersandGOFNMSmanagersandscientists.Thisdialog
onceagainhighlightedtheaforementionedbarrierstodataacquisition,visualizationanduse.
Toaddressthedataaccessbarrier,climatedatamanagersidentifiedseveralofNOAAsClimate
DataRecords(CDRs)thatwouldprovidereliable,defensible,andhomogeneousinformation
aboutenvironmentalvariablesthathadbeenidentifiedbythemarinemanagersandscientists
asimportanttoaddressingtheirclimaterelatedquestions.DirectaccesswasbuiltintoWCT
IMPACTviaanindexfilethateliminatedtherequirementthatauserknowwhere,andinwhat
formatthedataresidesbeforeitcanbeaccessed.Inadeparturefromcurrentdata
managementpractices,thevariablenamescontainedinthedatafile(e.g.,irwin)werereplaced
withthenamesthatmarinemanagersassociatewiththedata(e.g.,cloudcover).Toenhance
understandingofwhatdatawerebeingaccessed,whatcoveragewasavailable,andwhat
exactlywasbeingdescribedbythedata,NCDCclimatologistsalsoincludedaninformation
buttonthat,whenselected,displaysapopupwindowwithadescriptionofthedataandalink
tomorecompletemetadata.
Furthermore,WCTIMPACThasthefunctionalitytodisplaydataasanindividualtimesliceover
theareaofinterest,orifarangeoftimeisspecified,asananimation.Theanimationfeature
canbeespeciallyhelpfulforassistingamanagerindeterminingthedegreeofvariabilitythat
mayexistwithinthedata,ortoboundperiodswhenenvironmentalconditionsmayhave
exceededacriticalthresholdvalue.
AkeyaspectofWCTIMPACTisthatitisaJavatool,ratherthanaWebbasedtool.Thisallows
ittobeusedonacomputerthatisnotconnectedtotheInternet.Theadvantagesinclude
beingabletovisualizedatawhileworkinginthefield,ormodifyingananimationbasedon
feedbackfromastakeholderorpolicymakerduringabriefingataremotesite.WCTIMPACT
notonlyaccessesdatafromremoteservers,butalsocanaccessthedatastoredlocallyonthe
PC.Thus,adatasetmaybesubsettedoffNCDCsservers,storedasafileonalaptop,andthen
bereferencedwiththesoftwarewithoutafurtherInternetconnection.Thiscanbevery
importantformanagerswhohavelimitedorcostlyaccesstotheInternet.Locallycollected
datafilesmayalsobevisualizedinWCTIMPACTiftheyarestoredinarecognized
georeferencedformat.
VisualizationNeeds
Whenitcomestothevisualizationofthedata,ecosystemmanagerswereingeneralagreement
thattherewerecertainconsiderationsneededtoachieveaneffectivemarineecosystem
managementtoolforclimatedata.Firstofall,thetoolhadtobeaonestopshop.The
managerwantedaccesstotherelevantdatadirectlyfromthesoftware.Thisisachieved
throughtheWCTIMPACTdataaccesscode.Secondly,managerswantatoolthatdoesnt
33

requireasteeplearningcurve,orcomplexstepsinordertogettoavisualizationofthedata.
Theconsensuswasthatthetoolsinterfaceshouldhavealookandfeelthatdoesnotdiffer
dramaticallyfromtoolswithwhichtheyarefamiliar.Thetoolinterfaceshouldalsobesimple,
givinganoviceusertheabilitytointuitivelycreateabasicvisualization,butwithmorecomplex
optionsavailablethroughtabsormenus.
Currently,theWCTIMPACTtoolhasaseparatecontrolwindowandvisualizationwindow.A
complaintbymanagerswasthatonsmallerscreens(i.e.,laptop),onewindowwouldinevitably
obscuretheother,andwithmanywindowsopen,thecontrolwindowcouldbecomeburied
withnooptiontobringitforward.WCTIMPACTdevelopersarecurrentlyconsideringasingle
windowtabbedversionofthesoftwaresothatausermayquicklytransitionbetweenallactive
windowswithinthetool.
Managersalsostatedtheimportanceofbeingabletovisualizetwoormorevariablesinthe
samedisplay(e.g.,GISlayers).However,whenoverlaid,itbecomesdifficulttodifferentiate
colorschemesandpattersfromeachindividualvariable.WCTIMPACTdevelopersfirstthought
tocombineshadedreliefwithvectorcontours,howeversuchfunctionalityisacomplex
undertaking.Instead,itwasdecidedtoprovideamultiframewindow(Fig.2),whereuptofour
variablescouldbedisplayedsimultaneouslyandinsynchronization.Whenoneframeis
repositionedand/orzoomedinspace,orcycledintime,theothers,whenlinked,mirrorthe
alteration.

Figure3.FourframedatavisualizationwindowfromtheWCTIMPACTsoftware.

Inaddition,managersrequestedthatavisualizationtoolalsohavethecapacitytoperformat
leastbasicanalysesofthedata.Theseincludedthedisplayofsums,averages,differences,
34

anomalies,probabilities,extremes,andtrends.Furthermore,amanagershouldhavethe
optionofsavingthedisplaysalongwiththedatalocally,forexporttoothersoftware(e.g,
GoogleEarth)orforinsertingintoreportsandpublications(e.g.,briefingmaterials,factsheets).
Formanagers,oneofthemostcommonusesforvisualizationsoftwareistoprovidevisualaids
forexplainingaparticularphenomenontopolicymakers,stakeholders,orthepublic.
SUMMARY
TheWCTIMPACTdataaccessandvisualizationtoolhasbeendevelopedspecificallytoaddress
thestatedneedsofmarineecosystemmanagerstocombinedataaccessanddatavisualization,
anddeliverrelevantclimateinformationtosupportmanagementdecisions,develop
ecosystemscaleclimateassessments,andproducevisualaidsinsupportofeducationand
outreacheffortstothepublic,stakeholders,andpolicymakers.Ateverystepofdevelopment,
marineecosystemmanagershavebeenincludedintheframingofrequirements,sothatnot
onlyisthetooldevelopedtobeofoptimalusebythem,butthattheclimatedatamanagers
whomaintainthetoolalsobecomemoreawareoftheneedsofthemarineecosystem
community.Featuresareincludedbecausethemarinemanagershaveaskedforthem,rather
thanbecausethedatamanagersdesirethetooltoserveeverypossibleneedofeverypossible
user.WCTIMPACTiscurrentlystillinaprototypestage,butisexpectedtobereleased
operationallywithintheupcomingmonths,viaNCDCswebsite.

35

Contributions via Correspondence

TheMultiscaleIntegratedModelofEcosystemServices(MIMES)isamodelingframework
designedtoaddressthemagnitude,dynamics,andspatialpatternsofecosystemserviceflows,
values,andtradeoffs.Theoreticalprinciples,expertknowledge,anddataderivedrelationships
areusedtolinkdiversetypesofinformationaboutasystem.TheMIMESmodelcouplesthe
humanandnaturalsubsystemsbyconsideringtheinterdependenciespresentwithinand
betweenthenaturalandhumanspheresthatmakeupanecosystem.Inaddition,MIMES
providesananalyticalframeworkforunderstandingtheecosystemlevelconsequencesof
scenariosthatreflectspecificmanagementdecisions.Themodeloutputsareusedto
understandecosystemservicetradeoffsthatoccurinspaceandtimeasaresultofspecific
managementdecisionsorenvironmentalshifts(e.g.climatechange).MIMEShasbeenapplied
toanumberofcasestudies.Recenteffortshavebeenfocusedonunderstandingacoastaland
nearshoreenvironmentinnorthernMassachusettsthatencompassesalargeportionof
MassachusettsBayandStellwagenBankNationalMarineSanctuary.Forthiscasestudytwo
setsofresourcemanagementscenarioswereexplored.ForageFishscenariosconsiderthe
effectsofchangingcommercialfishingratesandtargetedspeciesofforagefishesincludingthe
impactsofopeningafisheryonsandlancespecies(notcurrentlytargetedbycommercial
fishing)andchangingratesoffishingonherringspecies(currentlytargetedbycommercial
fishingortakenasbycatch).Resultsdemonstratetradeoffsrelatedtowhalewatching,
commercialandrecreationalfisheriesandconservation.WindEnergyscenariosconsiderthe
developmentofcommercialandcommunityscalewindenergysitesandtheassociated
tradeoffsinvolvingwhalewatching,commercialandrecreationalfisheriesandconservation.
TheMIMESmodelprovidesanessentialanalyticalframeworkfromwhichtodeterminethe
outcomeofalternativedecisions,yetadditionaltoolsareneededtotranslatemodeloutputsto
adiversestakeholderaudience(includingmanagers)andsupportthedecisionmakingprocess.
TheMarineIntegratedDecisionAnalysisSystem(MIDAS)isavisualizationtooldesignedto
communicatetheMIMESmodelsresultsandsupportunderstandingofecosystemservice
tradeoffsrelatedtoalternativemanagementdecisions.MIDASisauserfriendly,webbased
interfacethatincorporatesfeaturesofopensourceGIS,participatorymapping,andsocial
collaboration.Overall,MIDASembracesaWeb2.0perspectiveincorporatingfeaturesof
participatoryinformationsharing,interoperability,usercentereddesign,andcollaboration.
ThroughMIDAS,userscandebatealternativepolicyoptions,proposenewmanagement
scenarios,andcontributetoanongoingdialogueaboutthefutureofcoastalandmarine
systems.PathsthroughMIDAScanbesharedbetweenuserswithsimilarordiffering
motivationshelpingtoidentifycompetitiveorcooperativeoutcomesandgenerallyfosteringan
appreciationforthedisparateperspectivesrelatedtohumanuseofmarineresources.

36

Formoreinformation,pleasecontactProf.LesKaufman([email protected])orProf.SuchiGopal
([email protected]).

Recommendations
Principles:
Theworkshopparticipantsexpressedinterestincontinuingthedevelopmentofwideraccessto
NOAAdatasets,communitystandardsonwaysofattributingcredittoresearcherssharingdata,
andamoreuniformapproachtothesummarizationofdatawithprivacyissues.
Theworkshopparticipantsrecommendedthedevelopmentofcollaborationswithsocial
scientistsandorcommunicationsexpertstodiscusshowthevisualizationofecosystemdata
variesaccordingtotheaudienceandthemediaavailabletothepublic.
Theworkshopparticipantsrecognizedthatdifferentaudiencesrequiredifferentinterfacesand
canprocessdifferentlevelsofcomplexity.
Itisessentialtoestablishadialogwiththestakeholdersandmanagersinordertounderstand
thespecificinformationthatisneeded.
Theworkshopparticipantsnotedthatthedevelopersofvisualizationtoolsshouldconsiderthat
theuserinterfaceshouldhaveeffectivetoolstoaidtheusersindeterminingwhetherthedata
presentedwillbeusefultothem.Thefactisthatwiththeadvanceddevelopmentofricher
onlineoptions,themorecomplexthesystemistouse,thegreateristheneedtoprovidehelp
text,videohelpfiles,andcarefulprogramming(suchasgooddefaultvalues)whichwill
facilitatetheuseoftheenduserinterface.
Theworkshopparticipantsidentifiedtheneedtoincludetheappropriatescaleofthe
ecosystem(spatialandtemporal)forthetypeofanalysesneeded,theinclusionoferrorand
uncertaintyrepresentationsintheproductsofferedtotheusersandtheinclusionofthehuman
factorinalltherepresentationsrelatedtotheecosystemvisualizationwerealsoidentifiedas
needs.
Theworkshopparticipantsidentifiedspecificcriteriathatcouldimprovethesuccessofthe
ecosystemdatavisualizationusedbythestakeholdersandmanagers.Thosecriteriaincluded
thesuggestiontobreakuptheecosystemcomponents/groupstohelpsimplifythegraphic
representation,standardizationofuseofcolorsinthevisualizations,theuseofcartoon
characterstohelpfacilitatetheunderstandingofthedata,theneedtoincludevisualanimation
oftheparametersvariabilityandoftheecosystemresponsetothatvariationinagraphic
manner,theneedtoguidetheaudiencethroughtheuseofhomologiestoestablish
37

connectionswithpersonallyfamiliarconcepts,andfinallytowalkthroughdataexplorationwith
themanagersandstakeholders.
Theworkshopparticipantsrecommendedthattheattributesofdatashouldincludedata
location,datatype,acquisitiondateandtime,dataandmetadataquality,andthemethodology
usedtocollectandprocessthedata.Thisshouldalsoincludeuncertaintywithinthedata,and
sourcesofpotentialerror.

Actions:
Theworkshopparticipantsrecommendthataworkinggrouponvisualizingspeciesdistribution
drivenbyclimatechangeandotherfactors,initiallyfocusingonfishes,composedofregional
scientistsandindividualswithpracticalknowledge,explorewaysofexpressingthekey
elementsofwhereresourcesspecieshavebeendistributedandwheretheywilllikelybe
distributedinthefuture.Theworkshopenvisionsthatthisworkinggroupwillworkmainlyvia
correspondenceutilizingwebbasedcommunicationsandmayoccasionalgatherforinperson
workshops.TheworkshopidentifiedChadKeith(NEFSC)asaleadinthiseffortandapointof
contactforthoseinterestedinparticipating.
Theworkshopparticipantsnotedthatdespitethecomprehensivecoverageandreachofthe
workshop,accomplishedwithmodestresources,manyNOAAofficesfromotherregionswere
unabletoparticipate,andlikewise,manyacademicandnonNOAAagencieswerenot
represented.TheworkshopencouragesotherNOAAregionalteamstoconsiderfollowon
workshopsbuildingonthethemesexploredinthisworkshopengagingotherregionsand
groups.
Currently,NOAAhasongoinginterest/usergroupsthatfocusonsubjectareasofcurrent
interesttoabroadaudiencebothwithinandoutsideNOAA.Theworkshopsuggeststhatan
ecosystemsdatavisualizationusersgroupsmaybetimelyandofwideinterestandtheideaof
developingsuchagroupbeexplored.ThisproposalmaybeofoverlappinginteresttotheGIS
usersgroup,whichshouldbecontactedtogaugeinterest.MichaelParkewasidentifiedasthe
leadtoexplorethiseffort.
Theworkshopparticipantsrecommendedtrainingbeprovided,specificallyrelatedtothe
developmentofapplicationsusingtheUnity3DgameengineandTerraViz,aswellasto
increasetheuseamongthescientificcommunityoftheIpythonenvironment.
TheworkshopencouragesScienceCommunicationandInformationTechnologytrainingfor
marinescientiststobetterutilizecurrentcapabilitiesoftheinformationage.Scientistshouldbe
informedonschemaofcolors,icons,representationsthatworkbestacrossabroadspectrum
ofaudiences.
38

Appendix 1. Agenda
Day1,February5
Start
Duration
(h:m)
(h:m)
9:00
9:15

0:15
0:30

9:45
10:15
10:45
11:00
11:30

0:30
0:30
0:15
0:30
0:30

12:00
12:30
13:30
14:00

0:30
1:00
0:30
0:30

14:30
15:00
15:15

0:30
0:15
0:30

15:45

0:30

16:15
16:45
17:00

0:30
0:15

Day2,February6
Start
Duration
(h:m)
(h:m)
8:00
0:05
8:05
0:30
8:35
0:30
9:05
0:30
9:35
0:30
10:05
0:15
10:20
0:30
10:50
11:20
11:50
12:20
13:20
15:00

0:30
0:30
0:30
1:00
1:40

AGENDA
WELCOME&INTRODUCTION
Changingperspectivesaboutachangingecosystem
Visualizingcomplexity:howcanweviewfoodwebandmultispeciesdataand
dynamicecosystemmodeloutputs
VisualizationandGamingofEcosystemModelScenarios
BREAK
Gettingmorefromdatathanjustwherethefishare
Visualizingthebaseofthefoodchain
Challenges,lessonslearned,andaFEWpotentialvisualizationsolutionsfor
synthesizingsciencetoinformManagementinsouthFlorida
LUNCH
Dynamicconceptualizationofhabitat
Designinggeospatialdatavisualizationsforageneralaudience
Dealingwithscientificandcommunicationuncertaintythroughdata
visualization
BREAK
NEFSCDataServer:access&visualizationwithPython
Webaccess,analysisandvisualizationofstandardizedoceanographicand
meterologicaldata
OceanographicdataanddataservicesatPacIOOS,aregionalassociationof
IOOS(remote5hours)
WRAP
END

AGENDA
ORIENTATION
Dataandvisualizationintegrationviawebbasedresources
VisualizingFitnessforPurpose
VisualizationsofHabitatandSitingDecisions
Habitatmodelingandvisualization
BREAK
Communicatinganunderstandingoflowertrophiclevelsinmarineeosystems
FindAndVisualizeNOAAData:AnOverviewoftheNOAAEarthInfomation
System(NEIS)andTerraViz(remote2hours)
Gettingtoobservingdatausigntheclimatologytoolkit
Developingaclimatedatavisualizationtoolformarineecosystemmanagers
LUNCH
WRAP
END

39

Annala,Johnand
Friedland,Kevin
Friedland,Kevin
Gaichas,Sarah
Gamble,Robert

Geis,Scott
Hyde,Kim
Kelble,Chris

Manderson,John
Butkiewicz,Tom
Townsend,Howard

Manning,Jim
Signell,Rich
Potemra,James

DiStefano,Massimo
Groman,Robert
Parke,Michael
Gallager,Scott

Pershing,Andy
Smith,Jeff
YoungMorse,Riley
Shein,Karsten

Appendix 2. Participants
Name
Annala,John
Babb,Ivar
Brady,Damian
Butkiewicz,Tom
Cannon,John
Carr,Ben
Chamberlain,Glenn
Cieri,Matthew
Cousins,Stephen
Curtis,Tobey
Darby,Forbes
DiStefano,Massimo
Friedland,Kevin
Gaichas,Sarah
Gallager,Scott
Gamble,Robert
Geis,Scott
GonzalezMirelis,Genoveva
Groman,Robert
Guida,Vincent
Hackathorn,Eric
Keith,Charles
Kelble,Chris
KimHyde
Klein,Emily
Langton,Richard
Manderson,John

Institution
GMRI
UConn
UMaine
UNH
NOAA
BU
NOAANMFS
MaineDNR
Umaine
NOAA
NOAA
WHOI
NOAANMFS
NOAANMFS
WHOI
NOAANMFS
NOAANMFS
IMR
WHOI
NOAANMFS
NOAAOAR
NOAANMFS
NOAAOAR
NOAANMFS
UNH
NOAANMFS
NOAANMFS

email
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]

40

Appendix 2. Participants continued


Name
Manning,James
Martin,MarieCaroline
Mayer,Larry
Moore,Peter
MorrisonRu
Orphanides,Chris
Parke,Michael
Pasquarella,Valerie
Pershing,Andy
Potemra,James
Ribera,Marta
SamuelChavez
Schmidt,Andre
Schroeder,Melanie
Shein,Karsten
Sheremet,Vitalii
Shyka,Tom
Signell,Rich
Smith,Jeff
Smith,Laurel
StaceBeaulieu
Stevens,Rachel
Townsend,Howard
Trinko,Tara
Vine,Norman
YoungMorse,Riley

Institution
NOAANMFS
NOAANMFS
UNH
MARACOOS
NERACOOS
NOAANMFS
NOAANMFS
BU
GMRI
PacIOOS
BU
UMass
ASA
NOAANESDIS
WHOI
NERACOOS
USGS
NOAAOAR
NOAANMFS
WHOI
NHFish&Game
NOAANMFS
NOAA
GMRI

email
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]

41

Appendix 3. Visualization Products and Services


Name

TerraViz

Organization Contact Phone Contact E-mail

NOAA/GSD 303.497.4437

NEIS (NOAA
Earth
Information
Services)
NOAA/GSD 303.497.6724

BCOBob Groman DMO/WHOI 508-289-2409

Sarah
Gaichas

Genoveva
GonzalezMirelis

NOAA
NMFS
Northeast
Fisheries
Science
Center,
Woods Hole 508 495 2016

IMR

+47 55236376

Data
Management

Julien.Lynge@noa
a.gov

Visualization Link to Product

Brief Description of Services


Provided

We create visualizations,
simulations, and games for
NOAA, DOE, DOD, and other
organizations.

Brief Product Description


TerraViz is a high-end,
3D visualization tool
being developed within
NOAA with the goal of
displaying nearly all
http://esrl.noaa.g NOAA data, past and
ov/neis/
future.
NEIS is a collection of
services to help users
find, access, analyze,
http://esrl.noaa.g and manage NOAA data,
past and future.
ov/neis/

Jebb.Q.Stewart@n
oaa.gov
X

[email protected] X

Sarah.Gaichas@n
oaa.gov

[email protected]

http://mapservice
.bcodmo.org/mapserv Geospatial access to
BCO-DMO managed
er/mapsol/index.php
datasets

Work with investigators to


manage their data from
proposal to preservation.
Collect sufficient metadata to
enable data reuse by others.
Metadata and data are
searchable, with
plotting/visualilzation options
available to aid in deciding
whether someone elses data
would be of use to you. Data
are accessible and can be
downloaded in many formats,

http://gephi.org/
and
network visualization
http://pajek.imfm. software--free and not
si/doku.php
produced by me

none

We provide distribution
maps and/or predicted
http://www.marea occurrence of benthic
no.no/en
fauna and seafloor habitat

Text-based access to
BCO-DMO managed
www.bcodmo.org datasets

42

You might also like