Using Widely Available Geospatial Data Sets to Assess the Influence

RESEARCH ARTICLE

s

Using Widely Available Geospatial Data Sets to Assess the Influence of Roads and Buffers on Habitat Core Areas and Connectivity Patrick Jantz "REN3CHOOLOF%NVIRONMENTAL 3CIENCE-ANAGEMENT 5NIVERSITYOF#ALIFORNIA 3ANTA"ARBARA #!

Scott Goetz1 7OODS(OLE2ESEARCH#ENTER &ALMOUTH-!

s

#ORRESPONDINGAUTHOR

SGOETZ WHRCORG  

.ATURAL!REAS*OURNALn

Volume 28 (3), 2008

ABSTRACTLand development pressures that threaten habitat core areas and connectivity are intensifying across the nation and extending beyond urbanized areas in the form of rural residential development. This is particularly true in the temperate forests of the northeastern United States. We used a suite of nationally available data sets derived from satellite imagery to identify core habitat areas of the northeastern United States, including impervious cover (urbanized and developed areas) and forest cover (canopy density). These were augmented with road network extent and density. We analyzed the influence of different types of unimproved roads and amount of forest cover on identification of the extent and configuration of roadless areas, and then assessed these core habitat areas in terms of land ownership (public, private) and management (parks, refuges, multi-use). We also derived patch connectivity metrics using a graph theory approach, making use of cost surfaces that accounted for the above variables and associated landscape metrics. A case study linking suitable habitat for a keystone predator is explored. Because increased conversion and fragmentation of many roadless areas by exurban development will exacerbate the likelihood of local species extinctions, and complicate efforts to preserve intact functional ecosystems, our results suggest a starting point for the construction of a more comprehensive and ecologically functional reserve network for the region. The use of widely available data sets demonstrated the capability for similar analyses to be conducted nationally or for other regions. )NDEXTERMSCONNECTIVITY CONSERVATION GRAPHTHEORY LANDSCAPEECOLOGY ROADLESSAREAS

INTRODUCTION (ABITATLOSSANDFRAGMENTATIONOFTHENATU RALLANDSCAPERESULTINGFROMURBANIZATION AREWIDELYKNOWNTOIMPACTVARIOUSASPECTS OFBIOLOGICALDIVERSITY(ANSENETAL  )NTHE5NITED3TATES MILLIONKMOF LANDWEREDEVELOPEDATURBANANDEXURBAN DENSITIES AS OF THE YEAR  "ROWN ET AL   ,OW DENSITY DEVELOPMENT IN PARTICULAR HASINCREASEDRAPIDLYANDISNOW THE DOMINANT DEVELOPMENT PATTERN IN THE 5NITED3TATES4HEOBALD)RWINAND "OCKSTAEL .EARLYOFALLHOUSING UNITSARECONTAINEDINAREASTYPIlEDBYLOW DENSITYDEVELOPMENTINAMATRIXOFNATURAL LANDCOVERTYPESIE THEWILDLANDURBAN INTERFACE 2ADELOFFETAL  4HE ROAD BUILDING THAT ACCOMPANIES DEVELOPMENT IS A MAJOR CONTRIBUTOR TO HABITATFRAGMENTATION7ATTSETAL  2OADS ALSO PRESENT BARRIERS TO WILDLIFE MOVEMENT AREADIRECTCAUSEOFMORTALITY TO WILDLIFE AND ACT TO INCREASE INTRODUC TIONSOFNON NATIVESPECIES-ADER &AHRIGETAL&ORMANAND!LEXANDER 0ARENDESAND*ONES'IBBSAND 3HRIVER #ONVERSELY ROADLESSAREAS HAVEHIGHERLEVELSOFNATIVEDIVERSITYAND FEWERINVASIVESPECIES'LENNONAND0ORTER  !SDEVELOPMENTCONTINUES WECAN EXPECT ROADLESS AREAS THOSE CORE HABITAT AREASRELATIVELYUNAFFECTEDBYHUMANDIS TURBANCE TODECREASEINSIZE NUMBER AND QUALITY(ANSENETAL !SARESULT THEYWILLBEEVERMOREDIFlCULTTOPROTECT PARTICULARLYINHISTORICALLYMOREDEVELOPED

REGIONSLIKETHENORTHEASTERN5NITED3TATES *ANTZETAL  3EVERAL ASSESSMENTS HAVE NOTED THE CON TRIBUTIONS THAT INVENTORIED ROADLESS AREAS )2!S COULDMAKETOTHECONSERVATIONOF BIODIVERSITYON53&OREST3ERVICE53&3 LANDS .OSS ET AL  3TRITTHOLT AND $ELLASALA,OUCKSETAL 4HE ADDITIONOF)2!STODESIGNATEDWILDERNESS AREASWOULDINCREASETHEREPRESENTATIONOF IMPORTANT BIOPHYSICAL CHARACTERISTICS IN THENATIONSRESERVENETWORKANDENHANCE PROTECTION OF THREATENED ENDANGERED OR IMPERILEDSPECIESONFEDERALLANDS$E6E LICEAND-ARTI,OUCKSETAL  "ECAUSE OF THE ECOLOGICAL CHARACTERISTICS OFROADLESSAREAS MUCHINTERESTHASBEEN FOCUSED ON PROHIBITING FURTHER ROAD CON STRUCTIONIN)2!S4HE53&3NETWORKOF )2!SACROSSTHE5NITED3TATESCOVERSABOUT  KM53&OREST3ERVICE  !TPRESENT ROADCONSTRUCTIONISALLOWEDIN ABOUTOFTHISAREA53&OREST3ERVICE  !LENGTHYREVIEWHASRESULTEDINTHE &OREST3ERVICERECOMMENDINGPROHIBITION OFADDITIONALCOMMERCIALLOGGINGANDROAD BUILDINGIN)2!S4URNER ALTHOUGH THEISSUEISPOLITICALLYCHARGED #ONNECTIVITY OF LANDSCAPE ELEMENTS CAN PLAYA CRITICALROLE INSPECIES PERSISTENCE AND ITS LOSS CAN LEAD TO LOCALIZED EXTINC TIONS LONG AFTER THE INITIAL DECREASE IN CONNECTIVITY 4ILMAN ET AL  #ARROLL ETAL #ONNECTIVITYOFCOREAREASIS ALSO IMPORTANT IN THE CONTEXT OF SPECIES RESPONSESTOCLIMATECHANGE WITHDISPERSAL Natural Areas Journal 261

PATHWAYS BETWEEN SUITABLE HABITAT AREAS NECESSARY TO ENSURE SPECIES VIABILITY ON LONGER TIME SCALES (ANNAH ET AL   $UE TO THE PACE OF LANDSCAPE CHANGE IN MANY AREAS IT IS IMPORTANT TO INCLUDE CHARACTERISTICS OF THE INTERVENING MATRIX WHEN ASSESSING CONNECTIVITY IE FUNC TIONALCONNECTIVITY 4HISWASNOTTHECASE HOWEVER INTHE53&3ASSESSMENTSNOTED EARLIER WHERENON FEDERALLANDSWERENOT INCLUDED4URNER )NTHECONTEXTOF CLIMATICDISRUPTION WIDESPREADPOLLUTION OF AIR AND WATERWAYS AND ALTERATION OF HISTORIC DISTURBANCE REGIMES FEW AREAS INTHEEASTCOULDBECONSIDEREDTOEXISTIN AhNATURALvSTATE DETERMINATIONOFWHICH WOULD REQUIRE A DETAILED ASSESSMENT OF CURRENT CONDITION IN RELATION TO HISTORICAL BASELINES (OWEVER ROADLESS AREAS ARE MORELIKELYTHANSETTLEDAREASTOMAINTAIN IMPORTANT ELEMENTS OF BIODIVERSITY AND TO BE MORE RESILIENT TO CONTINUED HUMAN DISTURBANCE-OREOVER INORDERTOASSESS THEPOTENTIALFORREMAININGROADLESSAREAS TOAUGMENTCURRENTCONSERVATIONNETWORKS IT IS NECESSARY TO DETERMINE THEIR EXTENT MANAGEMENTSTATUS ANDLANDSCAPECONlGU RATIONACROSSOWNERSHIPBOUNDARIES (ERE WE MAKE USE OF REMOTELY SENSED AND OTHER SPATIAL DATA SETS IN THE EASTERN 5NITED3TATESTO IDENTIFYFORESTEDROAD LESS AREAS HERE TERMED CORE AREAS UNDER DIFFERENT SCENARIOS  CHARACTERIZE CORE AREA MANAGEMENT AND  ANALYZE THE CONTRIBUTIONCOREAREASMAKETOLANDSCAPE CONNECTIVITYACROSSTHESTUDYAREAUSINGA GRAPHTHEORETICAPPROACH "ORROWINGFROM53&3ROADLESSAREACRITE RIA COREAREASINOURANALYSISHADTOBEAT LEASTHECTARES CONTAINNOIMPROVED ROADS ANDBENOCLOSERTHANMTOTHE NEAREST IMPROVED ROAD )N ADDITION WE EXPLOREDTHEINmUENCEOFMODIFYINGTHESE VARIABLES ANDTHEAMOUNTOFFORESTTREE COVER ONTHEDERIVATIONOFCOREAREASAND ASSOCIATED CONNECTIVITY 7E DISCUSS THE CONNECTIVITY ANALYSIS IN THE CONTEXT OF POTENTIALREINTRODUCTIONOFPRIMARYPREDA TORSTOTHEREGION7HILEOURANALYSISWAS RESTRICTEDTOTHEEASTERN5NITED3TATES THE APPROACH USED NATIONALLY CONSISTENT DATA SETS AND COULD WITH MINOR MODIlCATION BEEXTENDEDTOOTHERREGIONS

262 Natural Areas Journal

DATA SETS

Protected Area Database

Roads and Developed Areas 7EUSEDAROADSDATASETBASEDON  5NITED3TATES#ENSUS"UREAU4)'%2LINE &ILESENHANCEDBY'EOGRAPHIC$ATA4ECH NOLOGY )NC '$4 AND AUGMENTED WITH ROAD CLASSIlCATIONS BY %32) %NVIRON MENTAL3YSTEMS2ESEARCH)NSTITUTE 4ABLE  $EVELOPEDAREASWEREIDENTIlEDUSING A MAP DERIVED FROM $EFENSE -APPING 3ATELLITE 0ROGRAM $-30 /PERATIONAL ,INESCAN 3YSTEM /,3 IMAGERY WHICH HASTHEUNIQUECAPABILITYOFDETECTINGLOW LEVELS OF VISIBLE AND NEAR INFRARED RADI ANCEATNIGHT4HE$-30 /,3DATAWERE CONVERTEDTOPERCENTOFIMPERVIOUSSURFACE COVER AT  KM RESOLUTION FOR THE 5NITED 3TATES%LVIDGEETAL 7EANALYZED THEACCURACYOFTHESEMAPSINEARLIERWORK 'OETZAND*ANTZ  Tree Cover 4REECOVERINFORMATIONFORTHESTUDYAREA WASOBTAINEDFROMAGLOBAL MRESO LUTION CONTINUOUS lELDS TREE COVER MAP DERIVED FROM THE -ODERATE 2ESOLUTION )MAGING 3PECTRORADIOMETER -/$)3 IMAGERY (ANSEN ET AL  4ABLE   4HE MAP CONSISTS OF CONTINUOUS VALUES INDICATINGTHEFRACTIONOFEACHMPIXEL COVEREDBYTREESIE CANOPYDENSITY 

7EUSEDANATIONALPROTECTEDAREADATABASE $ELLA3ALAETAL CONSISTINGPRIMARILY OFSTATELEVEL'AP!NALYSIS0ROGRAM'!0 STEWARDSHIPDATACOMPILEDBYTHE#ONSER VATION "IOLOGY )NSTITUTE WHICH CLASSIlES LAND INTO ONE OF FOUR BROAD STEWARDSHIP CATEGORIES BASED ON THE LEVEL OF PROTEC TIONPARTICULAROWNERSHIPORMANAGEMENT REGIMESAFFORDBIOLOGICALRESOURCES3COTT ET AL  4ABLE   #ATEGORY  AND  LANDS ARE PROTECTED FROM CONVERSION AND ARE GENERALLY MANAGED FOR MAINTENANCE OFBIOLOGICALDIVERSITY%XAMPLESOFTHESE INCLUDE .ATIONAL 0ARKS 3TATE 0ARKS AND 53&37ILDERNESS!REAS#ATEGORYLANDS AREPROTECTEDFROMCONVERSIONBUTMAYBE SUBJECTTOUSETHATDEGRADESHABITATQUALITY SUCHASLOGGING-OST53&3LANDSFALLINTO THISCATEGORY#ATEGORYLANDSAREAFFORDED NOFORMALPROTECTIONS&INALLY WEUSEDA    53 .ATIONAL !TLAS 7ATER &EATURE DATA SET OF RIVERS STREAMS AND LAKES TO MASK OUT WATER FEATURES %32)   METHODS Scenario Development 4O BEGIN THE PROCESS OF IDENTIFYING CORE AREAS A LINE DENSITY FUNCTION WAS AP PLIED TO THE ROADS DATA SET 4HE FUNCTION CALCULATED THE DENSITY OF ROADS WITHIN A SPECIlEDSEARCHRADIUSFROMTHECENTEROF

4ABLE$ATESOFACQUISITION SCALERESOLUTION ANDVALIDATIONOFDATASETSUSED

Dataset

Date

Scale/Resolution

Accuracy

Roads

2000

1:100,000

Unassessed

Impervious Surface

1996 - 1997

1 km

Root Mean Squared Error of 11.30 (Elvidge et al. 2004)

Tree Cover

2001

500 m

Misclassification error of 11.5% (Hansen et al. 2002)

Protected Areas Database

1970 - 2006

1:100,000 or finer

Unassessed

Water Features

1995 - 2002

1:1,000,000

Unassessed

Volume 28 (3), 2008

EACHMXMPIXELOFANOVERLYING GRID4HISRESULTEDINAMRESOLUTION GRIDLAYERFORTHESTUDYAREA THEVALUESOF WHICH REmECT THE LINEAR DISTANCE OF ROAD INMETERSPERSQUAREKILOMETER WITHINTHE SPECIlED DISTANCE FROM THE PIXEL CENTER !  M CELL SIZE GRAIN WAS USED TO ACCURATELYCAPTURETHEDETAILOFCOREAREAS WHILEMINIMIZINGCOMPUTATIONALREQUIRE MENTS4HE SEARCH RADIUS WHICH WAS SET TOM FUNCTIONEDASABUFFERDEPTHSO THATANYROADWITHINTHISDISTANCEWOULDBE IDENTIlEDANDAVOIDEDINTHESELECTIONPRO CESS"UFFERDEPTHREFERSTOTHEMINIMUM DISTANCEFROMCOREAREAEDGETOTHENEAR ESTROAD&ORABASELINECASE UNIMPROVED ROADS GENERALLY DElNED AS ONE LANE DIRT ROADSNOTPASSABLEBYASTANDARDPASSENGER CAR WEREEXCLUDEDFROMTHEROADSDATABASE AND WERE THEREFORE NOT DETECTED BY THE DENSITY FUNCTION EFFECTIVELY PERMITTING THEIRPRESENCEINCOREAREAS &ROMTHERESULTINGGRIDS WATERBODIESWERE REMOVEDTOELIMINATESPURIOUSCOREAREAS ANDCELLSWITHADENSITYVALUEOFZEROWERE EXTRACTED!FTERAPPLYINGAFUNCTIONWHICH IDENTIlED DISCRETE CONTIGUOUS GROUPS OF CELLS GROUPSOFCELLSWITHAMINIMUMSIZE OFAPPROXIMATELYHECTARESWEREIDEN TIlED4HEMEANTREECOVEROFEACHAREAWAS THEN CALCULATED 4HOSE AREAS WITH VALUES LESSTHAN ACOMMONLYUSEDTHRESHOLD USED TO DISTINGUISH BETWEEN FOREST AND WOODLANDCOVERTYPES WEREREMOVEDFROM THEINITIALANALYSIS'ROSSMANETAL  &INALLY THEMEANIMPERVIOUSSURFACEAREA OFEACHCOREAREAWASCALCULATEDUSINGTHE IMPERVIOUSSURFACEMAP4HISWASDONETO VERIFY THAT THE IDENTIlED CORE AREAS WERE FREE OF STRUCTURES AND OTHER ELEMENTS OF THEBUILTENVIRONMENT &ORTHEBASELINESCENARIODESCRIBEDABOVE AMBUFFERDEPTHWASUSEDANDUNIM PROVEDROADSWEREALLOWEDINCOREAREAS &OURADDITIONALANALYSESWERECONDUCTEDTO EXPLORETHEEFFECTSOFBUFFERDEPTH UNIM PROVEDROADS ANDTREECOVERONOURIDEN TIlCATIONOFCOREAREAS BUFFERDEPTHS WEREINCREASEDFROMMTOM  UNIMPROVEDROADSWERENOTALLOWEDINCORE AREAS  BUFFER DEPTHS WERE INCREASED AND UNIMPROVED ROADS EXCLUDED AND  THETREECOVERTHRESHOLDREQUIREMENT WASREMOVED Volume 28 (3), 2008

!FTER IDENTIFYING CORE AREAS OWNERSHIP MANAGEMENT STATUS WAS SUMMARIZED FOR THE STUDY AREA IN ITS ENTIRETY AND ON A STATE BY STATE BASIS #ATEGORIES  AND  WERE COMBINED BECAUSE OF THEIR SIMILAR EMPHASISONPROTECTINGBIODIVERSITY7HILE THEDATASETDIDNOTINCLUDECOMPREHENSIVE DATAFORPRIVATELANDSINTHESTUDYREGION WEASSUMEDTHATSTEWARDSHIPINFORMATION FOR GOVERNMENTAL ENTITIES WAS COMPLETE ANDANYAREANOTOWNEDBYFEDERAL STATE COUNTY ORLOCALGOVERNMENTSWASPRIVATE ANDUNPROTECTEDANDWOULD THEREFORE BE CLASSIlEDASCATEGORYLANDS Landscape Configuration Metrics 4HENUMBER MEANAREA ANDMEANPERIM ETERAREARATIOOFCOREAREASWERECALCULATED FORTHEBASELINESCENARIO#OREAREAMETRICS WERECALCULATEDUSING&RAGSTATS-C'A RIGALETAL ONASTATEBYSTATEBASIS ASWELLASBYMANAGEMENTSTATUS #ONNECTIVITYWASCALCULATEDUSINGAGRAPH THEORETIC APPROACH WHICH MODELS THE LANDSCAPEASASETOFNODESPATCHES CON NECTEDBYEDGESPATHS THESETOFWHICH ARETERMEDAGRAPH-ATRIXOPERATIONSCAN BEPERFORMEDTOINVESTIGATEHOWINDIVIDUAL NODESCONTRIBUTETOLANDSCAPECONNECTIVITY ORHOWWELLCONNECTEDANENTIRELANDSCAPE IS 0ATH DISTANCE BETWEEN NODES CAN BE DElNEDAS%UCLIDEAN ORAFUNCTIONALDElNI TIONCANBEUSEDTHATINCORPORATESRELEVANT SURROGATESOFLANDSCAPEPERMEABILITYSUCH AS ROAD DENSITY AND FOREST COVER 'RAPH THEORYALSOPERMITSINCORPORATIONOFMORE DETAILEDDATASETSASTHEYBECOMEAVAILABLE AND CAN BE USED TO ASSESS THE EFFECTS OF SCALE5RBANAND+EITT5RBAN  4HEAVAILABILITYOFDETAILEDREMOTELYSENSED ANDOTHERSPATIALDATASETSMAKESTHEAP PROACHPARTICULARLYSUITABLEFORASSESSING CONNECTIVITY OVER LARGE AREAS #ORE AREA CONNECTIVITYWASAPPROACHEDFROMTHEPER SPECTIVEOFATHEORETICAL TERRESTRIALSPECIES WITHNODISPERSALTHRESHOLDANDNODElNED TIMEPERIODFORDISPERSAL/URASSUMPTION WAS THAT HIGHER FOREST COVER AREAS WOULD PROVIDE LESS EXPOSURE TO HUMAN CONTACT BEMORETRAVERSABLEFORTHETHEORETICALSPE CIES AND THAT HUMAN DISTURBANCE ROADS DEVELOPMENT AND AGRICULTURE AND WATER BARRIERSWOULDDECREASETRAVERSABILITY

&IRST ACOSTSURFACEWASCREATEDTOCALCULATE THEFUNCTIONALDISTANCEBETWEENCOREAREAS 4HECOSTSURFACEINCORPORATEDTREECOVER IMPERVIOUS SURFACE COVER ROAD DENSITY ANDWATERBODIES0ERCENTTREECOVERWAS SUBTRACTEDFROMTOCREATEANEWMAP WHERE CELLS WITH LOWER VALUES WOULD BE MORE COSTLY AND CELLS WITH HIGHER VALUES OFTREECOVERWOULDBELESSCOSTLYTOTRA VERSE3OTHATALLVARIABLESWEREWEIGHTED EQUALLY ROAD DENSITY WAS TRANSFORMED TO RANGEFROMTO4HETRANSFORMEDROAD DENSITYMAPWASTHENCOMBINEDWITHTHE IMPERVIOUS SURFACE AND TREE COVER MAPS CREATINGAMAPWITHVALUESBETWEENAND 7ATERBODIESWEREASSIGNEDAVALUEOF 4HISRESULTEDINACOMPOSITEINDEXOF HUMANDISTURBANCEANDTREECOVER )NTERPATCH CONNECTIVITY WAS ASSESSED WITH RECENTLY DEVELOPED SOFTWARE CALLED !RC2STATS WHICH USES A GRAPH THEORETIC APPROACH TO IDENTIFY LEAST COST PATHS BE TWEENHABITATPATCHESFROMWHICHNETWORK CONNECTIVITYMETRICSARECALCULATEDHTTP WWWNICHOLASDUKEEDUGEOSPATIALSOFT WARE!RC2STATSREQUIRESTWOINPUTS HABI TATPATCHESIE COREAREAS ANDACUMULA TIVEDISTANCESURFACE DERIVEDFROMTHECOST SURFACE THE VALUES OF WHICH INDICATE THE DISTANCEFROMAPARTICULARCELLTOTHENEAREST HABITATPATCH TAKINGTHECOSTOFTRAVERSING INTERVENINGCELLSINTOACCOUNT0ATCHLEVEL CONNECTIVITYMETRICSDERIVEDFROMLEASTCOST PATHSBETWEENPATCHESINCORPORATEASPECTS OFTHESURROUNDINGLANDSCAPE ALLOWINGFOR EVALUATION OF EACH PATCHS CONTRIBUTION TO DIFFERENT ASPECTS OF OVERALL LANDSCAPE CONNECTIVITY 4HE CUMULATIVE DISTANCE GRID WAS CONVERTED TO A TRIANGULATED IR REGULAR NETWORK 4). WHICH IS A THREE DIMENSIONALSURFACEREPRESENTATIONh(IGH ELEVATIONvAREASOFTHE4).CORRESPONDTO AREASMORECOSTLYFORATHEORETICALSPECIES TOTRAVERSE,EASTCOSTPATHSBETWEENCORE AREASWERECALCULATEDFROMEDGESANDNODES EXTRACTED FROM THE 4). "ECAUSE OF THE CONSIDERABLE COMPUTATIONAL REQUIREMENTS OFCALCULATINGLEASTCOSTPATHSFORTHESTUDY AREA THE COST SURFACE WAS DEGRADED TO MSPATIALRESOLUTIONANDAMINIMUM VERTICALDIFFERENCETHRESHOLDOFOFTHE MAXIMUM VERTICAL DIFFERENCE IN THE 4). WASSPECIlED^ FUNCTIONALDISTANCE UNITS  6ERTICAL DIFFERENCES LESS THAN THIS THRESHOLD WERE NOT RESOLVED IN THE 4). Natural Areas Journal 263

264 Natural Areas Journal

36,802

0 44,354 61,711 73,370 970,965 Total

0 0

0 0

0 0

0 0

0.1 25 0.1

0 2,838 Rhode Island

0

19,795 New Jersey

25

0

0.1 19 0.1 19 0.2 58 0.2 25,397 Maryland

59

0 0 0 0 0.2 28 0.5 61 12,895 Connecticut

0 0 Delaware

5,309

26

0.5

26

0.5

0

0

0.2

0.2 178

50 0.2

0.3

52

352 1

113

1

1,310 2

1 218

1,725

21,075

104,525 Kentucky

4,271 117,298 Pennsylvania

Massachusetts

1 1,242 1 1,498 3,605 4

3

2 2,523 2 2,875 3 4,429 4 127,605 North Carolina

4,933

1 1,216 2 1,887 3,323 103,515 Virginia

4,609

4

3

2

2 2,099

1,021 4

3 2,833

2,255 5

4

3,025

4,439

9

5 5,785

5,625 62,690

109,181 Tennessee

West Virginia

7 9,232 9 10,883 10 12,905 12 125,773 New York

14,620

6 1,565 24,950 Vermont

3,475

14

2,900

12

2,044

8

13 3,010 14 3,305 18 4,355 20 23,983 New Hampshire

4,783

17

Percent Area

14,645 19

Percent Area

16,353 25

Percent Area

No Unimproved Roads Unimproved Roads No Unimproved Roads

21,171 28

Percent Area

84,135

23,153

State Area

&OR THE BASELINE CASE  M BUFFER UN IMPROVED ROADS INCLUDED  DISCRETE CORE AREAS COVERING   KM WERE IDENTIlEDnAPPROXIMATELYOFTHESTUDY AREA&IGURE 4ABLE #OREAREASWERE PRESENT IN EVERY STATE WITH THE EXCEPTION OF2HODE)SLAND/NASTATEBYSTATEBASIS -AINEAND.EW9ORKCONTAINTHEGREATEST ABSOLUTE EXTENT OF CORE AREA 7HEN CORE AREAISCALCULATEDASAPERCENTOFSTATEAREA -AINEAND.EW(AMPSHIRERANKHIGHEST

State

Baseline Case

Maine

RESULTS

Unimproved Roads

"ECAUSE OF THE GREAT INTEREST GENERATED FORITSREINTRODUCTIONINTHE.ORTHEAST WE USED THE GRAY WOLF #ANIS LUPUS AS AN EXAMPLE OF A KEYSTONE SPECIES TO ASSESS CURRENT LANDSCAPE CONNECTIVITY BETWEEN HIGH QUALITY HABITAT AREAS IN UPPER .EW %NGLANDnFROM THE !DIRONDACKS OF .EW 9ORKTOTHE#ANADIANBORDER6ISUALINSPEC TIONOFLEASTCOSTPATHSGENERATEDSEPARATELY FORTHEREGIONREVEALEDCRITICALAREASWHERE CONNECTIVITYBETWEENCOREAREACLUSTERSWAS MAINTAINEDBYASMALLNUMBEROFPATCHES 7E BUFFERED THESE LEAST COST PATHS BY  KMONEITHERSIDECREATINGAKMWIDE CORRIDOR ENCOMPASSING hSTEPPING STONEv COREAREASWHICHCONNECTLARGERCOREAREA CLUSTERS7ECALCULATEDAREA PERCENTTREE COVER PERCENT IMPERVIOUS SURFACE COVER AND ROAD DENSITY FOR THE CORRIDOR "ASED ON THESE METRICS WE USED RESULTS FROM THELITERATUREONWOLFDISPERSALANDHABITAT REQUIREMENTS TO ASSESS THE SUITABILITY OF THISCORRIDORFORPROVIDINGSUITABLEHABITAT CONNECTIVITYINTHEREGION

4ABLE2OADLESSAREAKM INEACHSTATECONSIDERINGTHEINmUENCEOFDIFFERINGBUFFERZONEDEPTHSANDUNIMPROVEDROADS

Case Study of a Keystone Predator

500 m buffer

800 m buffer

&ROMTHELEASTCOSTPATHS THREEPATCHLEVEL CENTRALITYMETRICSWERECALCULATED THE FRACTIONOFSHORTESTPATHSTHATGOTHROUGH EACHPATCHBETWEENNESS  THEFRACTION OFPOSSIBLENODESCONNECTEDTOEACHPATCH DEGREE AND ONEDIVIDEDBYTHEAVER AGEDISTANCETOALLNODESFROMEACHPATCH CLOSENESS 7ESUMMARIZEDTHESEMETRICS BYMANAGEMENTSTATUSANDASSESSEDTHEIR SIGNIlCANCEFORLANDSCAPECONNECTIVITYAT DIFFERENTSCALES

Volume 28 (3), 2008

&IGURE2EMOTE ROADLESSAREASINTHEEASTERN5NITED3TATESIDENTIlEDUSINGAMDISTANCEFROMTHENEARESTROAD

4HEAREASAREDISTRIBUTEDHETEROGENEOUSLY ACROSSTHESTUDYREGION,ARGENUMBERSOF AREAS LIE IN NORTHERN AND WESTERN -AINE AND EXTEND SOUTH INTO .EW (AMPSHIRE WHERELARGECOREAREASARECONTAINEDWITHIN THE 7HITE -OUNTAINS .ATIONAL &OREST !N IRREGULAR STRIP OF CORE AREAS EXTEND ALONG THE NORTH SOUTH AXIS OF 6ERMONT MANY OF WHICH ARE CONTAINED WITHIN THE 'REEN -OUNTAINS .ATIONAL &OREST 3IG NIlCANT AREAS ARE ALSO CONTAINED WITHIN THE 'REAT 3MOKY -OUNTAINS .ATIONAL 0ARK IN 4ENNESSEE 4HE !DIRONDACK AND #ATSKILL3TATE0ARKSIN.EW9ORKCONTAIN RELATIVELY ISOLATED CLUSTERS OF CORE AREAS ASDOMOUNTAINOUSAREASOF0ENNSYLVANIA AND THE RIDGE AND VALLEY PHYSIOGRAPHIC PROVINCE 4HE REMAINING CORE AREAS ARE Volume 28 (3), 2008

SPARSELYDISTRIBUTEDACROSSTHESTUDYAREA MOSTLYALONGTHERIDGESOFTHE!PPALACHIAN -OUNTAINS 4HE THREE LARGEST AREAS ARE CONTAINED WITHIN THE !DIRONDACK 3TATE 0ARKOF.EW9ORK ,ANDSCAPE CONlGURATION METRICS DERIVED FROMTHEBASELINECASEVARIEDWIDELYFROM STATETOSTATEACROSSTHESTUDYREGION4ABLE   4HE NUMBER OF CORE AREAS PER STATE RANGEFROMJUSTONEIN$ELAWARETOIN -AINE3OMEAREASWERESPLITACROSSSTATE LINES WHICHINCREASEDTHENUMBERBY OVER THE  AREAS INITIALLY CALCULATED OVERTHESTUDYAREA-EDIANCOREAREASIZE RANGEDFROMKMIN-AINETOKM IN.EW*ERSEY

Alternative Derivations 5NIMPROVEDROADSHADALARGEEFFECTONTHE IDENTIlEDCOREAREAS&IGUREA %XCLUD ING AREAS CONTAINING UNIMPROVED ROADS DECREASEDTHEEXTENTOFCOREAREAWEIDENTI lEDBY#ONSIDERATIONOFUNIMPROVED ROADS HAD DIFFERENT EFFECTS ACROSS STATES 4ABLE A  &OR EXAMPLE 7EST 6IRGINIA EXPERIENCEDADECREASEINIDENTIlED CORE AREA WHILE .ORTH #AROLINA HAD A  DECREASE DESPITE COMPARABLE INITIAL AMOUNTS OF  KM FOR 7EST 6IRGINIA ANDKMFOR.ORTH#AROLINA4ABLE MBUFFER 5NIMPROVED2OADS  )NCREASINGTHEBUFFERDEPTHFROMMTO  M DECREASED THE EXTENT OF CORE AREA Natural Areas Journal 265

4ABLE#OREAREALANDSCAPEMETRICSFORTHEBASELINECASE BYSTATE

Baseline Case Number

Largest Patch

Median Area

Mean

Perimeter / Area

State

of Patches

(km2)

(km2)

Area (km2)

Ratio

Maine

290

826

45

80

13

61

249

42

57

18

142

1,932

40

103

19

70

466

34

68

22

Tennessee

122

264

32

47

19

North Carolina

110

533

29

45

22

Pennsylvania

132

107

29

32

19

West Virginia

156

235

29

36

24

Virginia

138

157

28

33

25

Delaware

1

26

26

26

13

Kentucky

77

61

23

22

33

Massachusetts

Vermont New York New Hampshire

13

33

20

17

27

Maryland

4

31

14

15

44

Connecticut

9

24

13

13

78

New Jersey

2

21

13

13

27

Rhode Island

0

0

0

0

0

IDENTIlED BY  AN EVEN LARGER EFFECT THAN THE EXCLUSION OF AREAS CONTAINING UNIMPROVEDROADS&IGUREB 4ABLEB  /N A STATE BASIS DECREASES IN THE EXTENT OFCOREAREAASBUFFERDEPTHINCREASEDCOR RESPONDEDROUGHLYWITHTHETOTALNUMBER OFCOREAREASINEACHSTATE

CATEGORYORLANDS 4ABLE 4HEGREAT EST PROPORTION OF CORE AREA  WAS IN CATEGORYLANDS.OTABLY ALMOSTOF COREAREASARESUBJECTEITHERTODEVELOPMENT ORMANAGEMENTACTIVITIESTHATCOULDMODIFY ORDEGRADEHABITATQUALITY

.EW9ORK 0ENNSYLVANIA AND.ORTH#ARO LINACONTAINEDTHEGREATESTAMOUNTOFFULLY PROTECTEDCATEGORYOR COREAREA4ABLE  !SAPERCENTAGEWITHINEACHSTATE .EW 9ORKAND.EW(AMPSHIRERANKEDHIGHEST ALTHOUGHABSOLUTECOREAREAIN.EW(AMP SHIRE WAS COMPARATIVELY SMALL -AINE WHICHHADTHEGREATESTOVERALLAMOUNTOF COREAREA HADONEOFTHELOWESTPROPORTIONS INTERMSOFPROTECTEDLANDS !NALYZING BY MANAGEMENT CATEGORY AR TIlCIALLY DIVIDED CONTIGUOUS CORE AREAS &OR EXAMPLE A SINGLE CORE AREA MAY FALL UNDERTHREEMANAGEMENTREGIMES)NTHIS CASE METRICSWERECALCULATEDONPORTIONS OFTHECOREAREAS WHICHREmECTOWNERSHIP CONlGURATION RATHER THAN ECOLOGICAL CON lGURATION #ATEGORY  LANDS WITH  CORE AREAS CONTAINED  MORE AREAS THANCATEGORYLANDSANDMORETHAN CATEGORY  AND  LANDS 4ABLE   -EAN COREAREAWASLOWESTFORCATEGORYLANDS ASWELL ATKM-EANAREAWASHIGHEST FORCATEGORYLANDSKM ANDLOWEST FORCATEGORYANDLANDSKM 0E RIMETERAREARATIOWASHIGHESTFORCATEGORY  LANDS AND LOWEST FOR CATEGORY  LANDS 4HESEVALUESWEREMUCHHIGHERTHANTHOSE CALCULATEDONASTATEBASIS 4HEEFFECTOFINCLUDINGUNIMPROVEDROADS INTHEDERIVATIONSOFWILDAREASWASMOST PRONOUNCED IN CATEGORY  LANDS WHICH DECREASEDBY2EDUCTIONSINCATEGORY  AND  LANDS WERE  VERSUS  IN CATEGORY  LANDS !S WITH UNIMPROVED ROADS DECREASING BUFFER DEPTH AFFECTED

2EMOVINGTHETREECOVERTHRESHOLDOF INCREASEDTHEAMOUNTOFCOREAREAIDENTIlED INTHESTUDYAREABYKM OVER THEBASELINECASE3OMESTATESEXPERIENCED MUCHGREATERINCREASESTHANOTHERS.ORTH #AROLINAAND-AINEINCREASEDMORETHAN  KM ALTHOUGH THE PROPORTIONAL IN CREASEIN.ORTH#AROLINAWASGREATER VERSUS   .EW *ERSEY HAD THE LARGEST PROPORTIONALINCREASEINCOREAREA CORRESPONDINGTOKM Management – Conservation Status 2OUGHLY  OF CORE AREA   KM IS CURRENTLY PROTECTED FROM DEVELOPMENT ANDHASSTRONGLANDUSECONTROLSIE '!0 266 Natural Areas Journal

&IGUREA %FFECTOFUNIMPROVEDROADSONROADLESSAREAIDENTIlCATION"LACKINDICATESAREASADDED WHEN UNIMPROVED ROADS ARE ALLOWED IN ROADLESS AREAS B %FFECT OF BUFFER DEPTH ON ROADLESS AREA IDENTIlCATION"LACKINDICATESAREASADDEDWHENBUFFERDEPTHISDECREASED

Volume 28 (3), 2008

40 40 29,016

24,910

0 0 0

0

28 26 3,737

3,673

31 29 6,800

6,525

31 31 1,479

1,345

46 41 1,432

1,335

43

53 2,340

1,905

51

42

2,952

2,058

63 59 2,722

2,108

66 60 3,371

2,003

66 65 2,773

2,363

67 68 40

39

56 76 166

63

86

100

80

25 100 25

1,374

1,131

100 100 26

26

100 28 100 61

Percent Area Percent Area

No Unimproved Roads Unimproved Roads

Increasing Buffer Depth from 500m to 800m

Total 17 16 Total

11,658

7,552

Rhode Island 0 0 Rhode Island

0

0

New York 0 0 Delaware

0

0

Maine 0 0 New Jersey

0

0

New Hampshire 0 2 Maryland

1

0

Vermont 10 9 Maine

1,983

1,708

Tennessee

North Carolina 9

12

9

352 10 504

New Hampshire

North Carolina

429

295

Virginia 15 12 New York

1,715

1,650

West Virginia 17 16 Pennsylvania

667

256

Pennsylvania 23 17 575 Vermont

478

Maryland 26 23 1,346 Tennessee

734

Massachusetts 49 24 416 Kentucky

174

New Jersey

Kentucky

55

36 671

1,233 46

28

2,601

1,286 Virginia

48 105 Massachusetts

53 32 Connecticut

Volume 28 (3), 2008

West Virginia

Delaware 3 2

Connecticut 0 0

Percent Area Percent Area State

800 m buffer 500 m buffer

Excluding Areas Containing Unimproved Roads

(b)

State

Connectivity and Case Study

(a)

4ABLEA #HANGEDECREASE INROADLESSAREAIDENTIlEDFOREACHBUFFERWIDTHAFTEREXCLUDINGUNIMPROVEDROADSB #HANGEDECREASE INROADLESSAREAIDENTIlEDFOREACHUNIMPROVEDROADS SCENARIOAFTERINCREASINGBUFFERDEPTH

CATEGORYLANDSMORETHANCATEGORYAND LANDS&ORTHEBASELINECASE INCREASING THEBUFFERDEPTHINCATEGORYANDLANDS DECREASEDCOREAREAIDENTIlEDBY)N CATEGORY  LANDS THE DECREASE WAS  4HELARGESTDECREASE  AGAINOCCURRED INCATEGORYLANDS)NCREASINGBUFFERDEPTH HAD A LARGER EFFECT WHEN CORE AREAS CON TAININGUNIMPROVEDROADSWEREEXCLUDED PRESUMABLY BECAUSE THE REMAINING AREAS WEREMOREFRAGMENTED-ODIFYINGTHETREE COVERTHRESHOLDLEDTOMINORCHANGESINTHE AMOUNTOFCOREAREAINEACHMANAGEMENT CATEGORY2EMOVINGTHETREECOVERTHRESH OLDOFINCREASEDTHEAMOUNTOFCORE AREA IDENTIlED IN CATEGORY  AND  LANDS BY INCATEGORYLANDSBY ANDIN CATEGORYLANDSBY

4HE BETWEENNESS METRIC WAS NOT COHER ENTLY DISTRIBUTED ACROSS THE STUDY AREA BUTAMAJORCORRIDORWASAPPARENTRUNNING NORTHSOUTHTHROUGH-AINEAND.EW(AMP SHIRE&IGURE INDICATINGHIGHLEVELSOF LANDSCAPE PERMEABILITY IN THE REGION ! SIMILAR ALTHOUGH LESS APPARENT CORRIDOR WAS EVIDENT IN 7EST 6IRGINIA -EAN BE TWEENNESSVALUESWEREHIGHESTFORCATEGORY ANDLOWESTFORCATEGORYLANDS #OREAREASINTHECENTRALPARTOFTHESTUDY AREA HAD HIGHER VALUES OF THE CLOSENESS METRIC INDICATING GREATER CLUMPING OF COREAREASTHANTOEITHERTHENORTHORSOUTH &IGURE #ATEGORYLANDSHADTHELOWEST CLOSENESSVALUES WHILEVALUESFORCATEGORY ANDASWELLASCATEGORYLANDSWERE COMPARABLE,ARGER CENTRALLYLOCATEDCORE AREASHADHIGHERVALUESOFTHEDEGREEMETRIC THANPERIPHERALLYLOCATEDAREAS INDICATING GREATER NODE CONNECTIVITY $EGREE VALUES WERE HIGHEST FOR CATEGORY  AND  AND LOWESTFORCATEGORYLANDS 7E IDENTIlED A  KM CORRIDOR CON NECTING LARGE CORE AREAS IN !DIRONDACK 3TATE 0ARK WITH THOSE IN 6ERMONT AND .EW(AMPSHIRE&IGURE 4HECORRIDOR CONTAINS HIGH TREE COVER  BUT RELA TIVELYHIGHROADDENSITIESKMKM THATMAYIMPEDEDISPERSALOFSPECIESLIKE WOLVES)MPERVIOUSSURFACECOVERWASLOW ALTHOUGHTHECORRIDORISPRIMARILYUNPRO

Natural Areas Journal 267

4ABLE-ANAGEMENTSTATUSOFROADLESSAREASKM CONSIDERINGTHEINmUENCEOFDIFFERINGBUFFERZONEDEPTHSANDUNIMPROVEDROADSSEESECTION 

500 m buffer Unimproved Roads

800 m buffer

No Unimproved Roads

Unimproved Roads

No Unimproved Roads

GAP Status

Area

Percent

Area

Percent

Area

Percent

Area

Percent

Category 1 or 2

17,178

23

15,562

25

12,645

29

11,233

31

Category 3

30,831

42

27,219

44

20,238

46

17,596

48

Category 4

25,360

35

18,931

31

11,471

26

7,973

22

Total

73,370

100

61,711

100

44,354

100

36,802

100

TECTEDIN'!0CATEGORY ANDTHUS SUBJECTTOPOSSIBLEDEVELOPMENT DISCUSSION Core Areas, Management Status, and Alternative Derivation Scenarios )N THE BASELINE SCENARIO   KM OF CORE AREA WERE IDENTIlED IN THE EASTERN 5NITED 3TATES &OR PERSPECTIVE THERE IS A TOTAL OF  KM OF 53&3 INVENTORIED ROADLESSAREASINTHESTUDYAREAANDATOTAL OF KMOFLANDIN'!0CATEGORY ORSTATUS4HUS COMPAREDTOTHECURRENT RESERVE NETWORK THERE IS A CONSIDERABLE AMOUNTOFROADLESS FORESTEDCOREHABITATIN THEEASTERN5NITED3TATES4HEEXTENTOFCORE AREAIDENTIlEDISSENSITIVETOASSUMPTIONS MADEABOUTDIFFERENTROADTYPESANDHOWFAR THEIRINmUENCEEXTENDSFROMTHEROADEDGE )NOURMOSTRESTRICTIVESCENARIO USINGAN MBUFFERDEPTHANDALLOWINGNOUNIM PROVEDROADS THEEXTENTOFAREAIDENTIlED DROPPEDBYHALFTO KM  4HE EXTENT OF IDENTIlED CORE AREA WAS ALSODEPENDENTONTHEDATASETSUSED!TA SCALEOF ITISPOSSIBLETHATROADS WEREMISSEDINTHEDIGITIZINGPROCESS)NA COMPARISONOFROADSDATASETSFOR.ORTHERN 7ISCONSIN (AWBAKERAND2ADELOFF FOUND THAT ROAD DENSITIES CALCULATED US ING THE 4)'%2  DATA SET WERE LESS THAN HALF THOSE CALCULATED FROM DIGITIZED $/113 WITHMOSTOFTHEDIFFERENCEDUE TO OMISSION OF SMALLER ROADS )T IS LIKELY THAT MANY SMALL PAVED AND UNIMPROVED ROADS PRESENT IN OUR STUDY AREA WERE NOT REPRESENTED IN THE DATA SET WHICH WOULD LEADTOANOVERESTIMATEOFTHEEXTENTOFCORE 268 Natural Areas Journal

AREASANDANUNDERESTIMATEOFTHEEFFECTOF UNIMPROVEDROADSONCOREAREAIDENTIlCA TION)NTHEABSENCEOFREGIONALACCURACY ASSESSMENTSFORTHE4)'%2DATASET THE MAGNITUDE OF THE EFFECT IS CURRENTLY DIFlCULTTOASSESS ALTHOUGHMOSTOMITTED ROADSWOULDBEUNIMPROVEDADDRESSEDIN OURSCENARIOASSESSMENT  4HE MANAGEMENT REGIME FOR CORE AREAS VARIES CONSIDERABLY /VER A THIRD OF CORE AREASAREINMANAGEMENTREGIMESTHATOF FERNOSPECIlEDLEGALPROTECTION!NOTHER  ARE PROTECTED FROM CONVERSION BUT SUBJECTTOACTIVITIESTHATDEGRADETHEQUALITY OFBIOLOGICALRESOURCESEG COMMERCIAL LOGGING 4HUS CLOSETOOFTHECORE AREASIDENTIlEDARESUBJECTTOUSESWHICH MAYDECREASETHEIRCONTRIBUTIONTOCONSER VATION OF BIODIVERSITY 4HESE PROPORTIONS VARY FROM STATE TO STATE -OST OF .EW 9ORKSCOREAREA WASINCATEGORY ORLANDS WHEREASOF-AINESWAS IN CATEGORY 7EST6IRGINIAS CORE AREAS WEREALMOSTEVENLYSPLITBETWEENCATEGORY ANDLANDS WITHONLYASMALLFRACTION MORE FULLY PROTECTED "ECAUSE PRIVATE LANDSWERENOTSYSTEMATICALLYSURVEYEDFOR

DEVELOPMENTOFTHE0ROTECTED!REA$ATA BASE,OUCKSETAL WEHAVELIKELY UNDERESTIMATEDTHEEXTENTOFPROTECTEDCORE AREAONPRIVATELANDS 2EMOVING THE  TREE COVER THRESHOLD ASACRITERIONFORINCLUSIONASACOREAREA HAD LESS EFFECT ON THE ALTERNATIVE DERIVA TION SCENARIOS THAN EXPECTED INCREASING THEAREAIDENTIlEDBYJUST4HISRESULT INDICATESTHATRELATIVELYFEWOFTHEIDENTIlED AREASWEREOPENVEGETATIONTYPES SUCHAS MARSHES WHERE TREE COVER IS PERENNIALLY LOW#ONVERSELY ITINDICATESTHATMOSTOF THE CORE AREAS IN THE %AST ALREADY HAVE HIGH DENSITIES OF TREE COVER THUS WOULD UNDER MOST DElNITIONS BE DESCRIBED AS FOREST ALTHOUGH THE CONDITION OF THESE FORESTED AREAS LIKELY VARIES BASED ON AL LOWABLE LAND USES /N CATEGORY  AND  LANDS IN PARTICULAR ONE MIGHT EXPECT TO lND EVIDENCE OF RECENT ANTHROPOGENIC DISTURBANCE $EPENDING ON THE SCALE AND TYPEOFDISTURBANCEEG WHETHERCLEARCUT LOGGINGORSELECTIVEHARVESTING THEIMPACT MAYNOTBEGREATENOUGHTODECREASEAVER AGECOVERINAPARTICULARCOREAREABELOW THE  THRESHOLD 4HIS IS TRUE IN 7EST

4ABLE#OREAREAMETRICSFOREACH'!0LANDMANAGEMENTCATEGORY

Baseline Scenario Gap Category

Number of Patches

Mean Patch Area (km2)

Median Patch Area (km2)

Mean Perimeter Area Ratio

1&2

1,180

15

0.9

75

3

1,786

17

1.4

68

4

6,796

4

0.2

100

Volume 28 (3), 2008

0

0

0

0

0

0

0

0 0

0 0

0

0

0

Delaware

Rhode Island

0

24

100

0

15 0 0 0 7 Massachusetts

6

103

79

19

0 0 0 0 6 Connecticut

59

5

41

0

1 0 0 0 15 Maryland

34

28

65

0

0 0 0 0 17 New Jersey

77

5

23

0

24 1 0 0 123 Kentucky

15

482

61

188

45 1

17 1 592

2

44

2

1,389 1 323 West Virginia

10

1 783 Virginia

23

2,052

60

1,422

37 2 1 1 1,107 Tennessee

29

1,321

34

1,436

5 8 24 1 1,104 Maine

5

19,856

90

1,034

14 1 1 1 1,817 North Carolina

44

1,737

42

601

35 1 5 2 Vermont

405

15

1,308

50

912

24 4 0 2 Pennsylvania

2,144

72

148

5

705

29 1,119

28 4

21

3,443 3

9

53

0 325

2,050 18 3 713

70 7 8,614 New York

New Hampshire

Percent of Roadless Area Percent of State Area State

Area

Percent of State Area

Percent of Roadless Area

Area

Percent of State Area

Percent of Roadless Area

Area

Category 4 Category 3 Category 1 or 2 GAP Status

4ABLE-ANAGEMENTSTATUSOFROADLESSAREASBYSTATE FORTHEBASELINESCENARIO SHOWINGPERCENTOFROADLESSAREAKM  UNDEREACHMANAGEMENTTYPE'!03TATUS 

Volume 28 (3), 2008

6IRGINIA FOR EXAMPLE WHERE SELECTIVE HARVESTING OF VALUABLE HARDWOODS LIKE CHERRY 0RUNUS SPP AND OAK 1UERCUS SPP HAS ACCELERATED IN RECENT YEARS AND WHEREINCLUSIONOFUNIMPROVEDROADSINOUR DATABASEISLIKELYUNDERESTIMATED !NOTHERISSUEARISESWHENMANAGEMENTAL TERSSERALSTAGEORFORESTTYPE&OREXAMPLE NATIVEHARDWOODFORESTSONTHE#UMBERLAND 0LATEAU IN 4ENNESSEE ARE BEING CLEARCUT ANDREPLANTEDWITHNON NATIVEPINES-C 'RATHETAL /URANALYSISDOESNOT DISTINGUISHBETWEENNON NATIVEPLANTATIONS AND NATIVE FOREST STANDS ALTHOUGH THEY FUNCTIONDIFFERENTLYECOLOGICALLYANDCOULD SIGNIlCANTLYALTERTHECONSERVATIONVALUEOF AFFECTEDCOREAREAS )NLIGHTOFTHEVARIETYOFDATASETSUSEDIN THISSTUDY ATTENTIONTOTHEINmUENCEOFSCALE WASREQUIRED!LLDATASETSWEREREFERENCED TO A MASTER EXTENT AND ALL CALCULATIONS WEREMADEUSINGAMCELLSIZE4HIS REQUIREDASSUMPTIONOFAUNIFORMWITHIN CELL DISTRIBUTION OF TREE AND IMPERVIOUS COVER AND RESULTED IN SUPERSAMPLING OF THECOARSERRESOLUTIONDATA(OWEVER COVER VALUESWEREAVERAGEDTHEREFORE WEEXPECT ANYSUPERSAMPLINGEFFECTSTOBESMALLAND PRIMARILYATTHEEDGESOFCOREAREAS 5SINGARESOLUTIONOFMTOCREATETHE ROADDENSITYDATASETALLOWEDUSTOCAPTURE lNER DETAILS OF THE SPACE BETWEEN ROADS )N SOME CASES THIS RESULTED IN IRREGULAR CORE AREAS WITH SPURS AND hBOTTLENECKSv 4HERESOLUTIONOFTHEIMPERVIOUSSURFACE ANDTREECOVERMAPSCOULD INSOMECASES BELARGERTHANTHESESPURSANDBOTTLENECKS "ECAUSEWECALCULATEDTREEANDIMPERVIOUS SURFACECOVERATTHEPATCHLEVEL WEDONOT EXPECTSYSTEMATICERRORINOURESTIMATESOF COVERFORCOREAREAS (IGHER RESOLUTION IMPERVIOUS AND TREE COVER DATA SETS DERIVED FROM ,ANDSAT  MIMAGERYARECURRENTLYAVAILABLEFORTHE COTERMINOUS53 ALTHOUGHWEHESITATETO SPECULATE HOW OUR RESULTS WOULD CHANGE USINGHIGHERRESOLUTIONDATA(IGHERSPATIAL RESOLUTION ENABLES IDENTIlCATION OF lNER SCALE FEATURES IN THE LANDSCAPE BUT THE SPECTRALRESOLUTIONOF-/$)3ISSUPERIOR TO ,ANDSAT AND THE DIFlCULTY OF DETECT ING LOW LEVELS OF IMPERVIOUSNESS WITH Natural Areas Journal 269

&IGURE4HE@BETWEENNESSCONNECTIVITYMETRICFORAPORTIONOFTHESTUDYAREA FOCUSEDON.EW%NGLAND WITHACUMULATIVECOSTSURFACEDESCRIBEDINSECTION  SHOWN IN THE BACKGROUND GRAYSCALE  "ETWEENESS VALUES RANGE FROM YELLOWS LOW TO BLUES HIGH 4HE HIGH VALUES RUNNING THROUGH THE CENTER OF THE REGIONINDICATEAHIGHDENSITYOFLEASTCOSTPATHSTRAVERSINGTHOSEHABITATPATCHES4HEOUTLINEDAREADEPICTSTHATANALYZEDIN&IGURE

270 Natural Areas Journal

Volume 28 (3), 2008

Volume 28 (3), 2008

Natural Areas Journal 271

&IGURE  4HE @CLOSENESS CONNECTIVITY METRIC FOR A PORTION OF THE STUDY AREA FOCUSED ON THE CENTRAL!PPALACHIANMOUNTAINSOF0ENNSYLVANIA WITHTHEBACKGROUNDCOSTSURFACEGRAYSCALE  6ALUESRANGEFROMYELLOWTOBLUEASPERlGURE4HECOREAREASINTHISREGIONHAVEHIGHVALUES REmECTINGTHEIRKEYROLEASNODESTHATHAVETHEPOTENTIALTOCONNECTHABITATSOF.EW%NGLANDTO THESOUTHERN!PPALACHIANS

&IGURE2EGIONALANALYSISOFCONNECTIVITYBETWEENFORESTSINTHE!DIRONDACKSOF.EW9ORKTO CORE HUBS IN6ERMONT AND .EW (AMPSHIRE4HE @DEGREE CONNECTIVITY METRIC IS SHOWN WITH VALUESRANGINGFROMYELLOWSLOW TOBLUESHIGH ANDABACKGROUNDCOSTSURFACEGRAYSCALE  -AJORLEASTCOSTPATHSAREOVERLAIDASREDLINES SHOWINGPRIMARYCONNECTIVITYROUTESDERIVED USINGTHEGRAPHTHEORYAPPROACH!CORRIDORCROSSHATCHING CENTEREDONLEASTCOSTPATHSWAS IDENTIlEDASCRITICALFORMAINTAININGCONNECTIVITYBETWEENLARGERCLUSTERSOFCOREAREAS

THE ,ANDSAT SENSOR ARE WELL KNOWN 4HE hNIGHTTIMELIGHTSvDATASETMAY THEREFORE BEMOREAPPROPRIATEFORDETECTINGDIFFUSE LOW DENSITY DEVELOPMENT LIKELY TO OCCUR NEAR CORE AREAS AND AT THE SCALE OF THIS STUDY THE-/$)3DATAAREWELLSUITEDTO CONSISTENTLY MAPPING LANDSCAPE PATTERNS OFCANOPYDENSITY Landscape Context and Connectivity

PERIMETER AREA RATIO VALUES INDICATING RELATIVELYREGULARPATCHEDGESCOMPAREDTO THOSEIN7EST6IRGINIA FOREXAMPLE4HE HIGHESTPERIMETERAREARATIOVALUESWEREIN STATESWITHRELATIVELYLITTLECOREAREA4HERE WEREINTERESTINGDIFFERENCESBETWEENMAN AGEMENTCATEGORIES HOWEVER#ATEGORY LANDSGENERALLYCONSISTEDOFALARGENUMBER OF SMALL PATCHES LEADING TO HIGH PERIM ETERAREARATIOS5NEXPECTEDLY CATEGORY LANDSCOMPRISEDLARGERPATCHESWITHLOWER PERIMETERAREARATIOSTHANCATEGORYAND LANDS4HISWASLIKELYDUETOTHELARGE EXTENSIVECOREAREASIN-AINE

-OST CORE AREAS WERE CLUSTERED IN THE NORTHEASTERN PORTION OF THE STUDY AREA WITH -AINE ALONE CONTAINING  OF THE TOTAL -AINES CORE AREAS WERE ALSO THE LARGEST WITHAMEDIANSIZEOFKM4HE LARGESTCOREAREAINTHESTUDYAREAWASIN !DIRONDACK 3TATE 0ARK  KM  4HE EFFECT OF LARGE CORE AREAS ON STATISTICAL SUMMARIESBYSTATECANBESEENINTHEDIF FERENCEBETWEENMEDIANANDMEANAREAS &OREXAMPLE MEDIANCOREAREASIZEIN.EW 9ORK WAS  KM WHILE MEAN CORE AREA SIZEWASKM)NGENERAL THELARGEST COREAREASWEREASSOCIATEDWITHNATIONALOR STATEPARKSANDFORESTS!NEXCEPTIONWERE THELARGEINDUSTRIALFORESTSOF-AINE WHICH HAVE UNDERGONE TREMENDOUS CHANGES IN OWNERSHIPOVER KMINTHEPERIOD TO ANDPARCELIZATIONTHATWILL DETERMINE THE FUTURE OF CONSERVATION IN THEREGION(AGANETAL 4HERECENT TREND TOWARDS MORE OWNERS AND SMALLER PARCELS IN -AINE AND OTHER STATES ACROSS THE%ASTERN53WHERELARGETIMBERCOM PANIESAREDIVESTINGTHEIRLANDSISLIKELYNOT REmECTEDINOURDATASETS&OREXAMPLE REAL ESTATE INVESTORS PART OF THE NEW ARRAY OF LANDOWNERSIN.ORTHERNFORESTS MAYHOLD LAND FOR SEVERAL YEARS BEFORE CONDITIONS AREPROlTABLEFORDEVELOPMENT WHEREUPON DECREASESINCANOPYCOVERANDINCREASESIN IMPERVIOUSNESS WOULD BECOME APPARENT IN REMOTELY SENSED DATA )N MANY CASES NEW OWNERS ARE LESS LIKELY TO ENGAGE IN ACTIVITIESTOPROMOTEBIODIVERSITY.EGATIVE EFFECTS ON HABITAT AND BIODIVERSITY STEM MINGFROMSHIFTSTOWARDSLESSSUSTAINABLE MANAGEMENT EG DECLINING TO OBTAIN FOREST CERTIlCATION HAVE BEEN OBSERVED (AGANETAL 

/URANALYSISOFCOREAREACONNECTIVITYIN THENORTHEAST USINGTHEGRAYWOLFASAFOCAL SPECIES ILLUSTRATEDTHEUTILITYOFTHEGRAPH THEORETICAPPROACHFORCONSERVATIONPLAN NING /THER APPLICATIONS OF THE APPROACH HAVEFOCUSEDONTHEORETICALSPECIES"UNN ETAL ANDCARIBOU2ANGIFERTARAN DUS /"RIENETAL 4HE7ILDLANDS 0ROJECT AND $EFENDERS OF 7ILDLIFE HAVE IDENTIlEDFORESTEDAREASIN-AINEANDTHE !DIRONDACKSSUITABLEFORSUPPORTINGWOLF POPULATIONSBUTFUTUREDEVELOPMENTCOULD REDUCESUITABILITYIFCONNECTIVITYBETWEEN CORE AREAS DECREASED IMPAIRING THE ABIL ITY OF LARGE AREAS SUCH AS !DIRONDACK 3TATE0ARK TOSUSTAINVIABLEWOLFPOPULA TIONS WITHOUT DISPERSAL FROM AREAS TO THE NORTHEAST

.OSTRONGPATTERNWASFOUNDINPERIMETER AREA RATIO VALUES AT THE STATE LEVEL #ORE AREAS IN -AINE HAD SOME OF THE LOWEST

7E IDENTIlED ONE POSSIBLE CORRIDOR FOR WOLVESDISPERSINGBETWEENFORESTSIN-AINE ANDTHE!DIRONDACKS2OADDENSITYINTHE

272 Natural Areas Journal

#ONNECTIVITY MEASURES VARIED WIDELY ACROSS THE STUDY AREA BETWEEN METRICS "ETWEENNESS VALUES WERE HIGHEST FOR A LINEARSETOFCOREAREASIN-AINEAND.EW (AMPSHIRE INDICATING LOW RESISTANCE TO DISPERSAL IN THE REGION #LOSENESS VALUES WEREHIGHESTINTHECENTRALPARTOFTHESTUDY REGION APATTERNSIMILARTOWHATONEWOULD EXPECTWHENMEASURINGAVERAGE%UCLIDEAN DISTANCEBETWEENCOREAREAS$EGREEVALUES WERE HIGHEST FOR LARGE CENTRALLY LOCATED COREAREAS3OMECOREAREASINTHE.ORTH EASTHADBOTHHIGHBETWEENNESSANDDEGREE VALUES INDICATINGTHEIRIMPORTANCEFORBOTH LOCALANDREGIONALCONNECTIVITY Case Study of a Keystone Predator

CORRIDOR IS HIGHER THAN THAT TOLERATED BY WOLVES IN THEIR HOME RANGES FOR WHICH THE UPPER THRESHOLD RANGES FROM  TO  KMKM 4HIEL  -ECH  -LADENOFFETAL (OWEVER ADJACENT AREASOFHIGHQUALITYHABITATCANINCREASE THETOLERATEDROADDENSITYTO^KMKM &ULLERETAL ANDDISPERSINGWOLVES WILLTRAVELACROSSCONSIDERABLEDISTANCESOF UNSUITABLEHABITATINCLUDINGMAJORROAD WAYS TOREACHFAVORABLEHABITAT-ECHET AL .ONETHELESS THEPROBABILITYOF SUCCESSFUL DISPERSAL DECREASES AS HUMAN ASSOCIATEDMORTALITYRISKINCREASES3HEP HERDAND7HITTINGTON 4HEHIGHTREE COVERANDLOWLEVELOFDEVELOPMENTINTHE CORRIDORWEIDENTIlEDINDICATESLIMITEDDIS PERSALMAYCURRENTLYBEPOSSIBLE ALTHOUGH PROTECTIONANDRESTORATIONOFCRITICALAREAS WOULD INCREASE THE LIKELIHOOD OF ITS SUC CESS4HISEXAMPLEREVEALSASPECTSOFTHE LANDSCAPE IMPORTANT FOR WOLF DISPERSAL AND THE APPROACH COULD BE TAILORED TO OTHER SPECIES OR SYSTEMS &OR EXAMPLE RIPARIAN CORRIDOR CONNECTIVITY COULD BE MEASUREDUSINGHIGHLYDETAILEDLANDCOVER MAPSINCORPORATINGVEGETATIONDENSITYAND STRUCTUREWITHINSPECIlEDBUFFERDISTANCES FROMSTREAMS'OETZ  CONCLUSION $EVELOPMENTPRESSURESTHATTHREATENCORE HABITAT AND CONNECTIVITY ARE INTENSIFYING IN MUCH OF THE NATION AND EXTENDING FAR BEYOND URBANIZED AREAS IN THE FORM OF RURALRESIDENTIALDEVELOPMENTANDSECOND HOMEBUILDING3OMEOFTHEHIGHESTRATES OF EXURBAN DEVELOPMENT IN THE COUNTRY AREOCCURRINGIN%ASTERNTEMPERATEFORESTS PARTICULARLY .EW9ORK .EW (AMPSHIRE 6ERMONT AND -AINE )F CURRENT TRENDS CONTINUE INCREASED FRAGMENTATION AND CONVERSION BY EXURBAN DEVELOPMENT OF MANY OF THE CORE AREAS WE IDENTIlED IS LIKELY PARTICULARLY BECAUSE  REMAIN UNPROTECTED .OT ONLY WOULD THIS BE AN IMPORTANTLOSSOFHABITATAREA BUTWOULD FURTHER FRAGMENT AND ISOLATE HABITAT IS LANDS DECREASINGSUCCESSFULDISPERSALAND INCREASINGTHELIKELIHOODOFLOCALSPECIES EXTINCTIONS /UR RESULTS SUGGEST A STARTING POINT FOR THECONSTRUCTIONOFAMORECOMPREHENSIVE

Volume 28 (3), 2008

RESERVENETWORKFORTHESTUDYREGIONAND CANALSOAIDBROADERECOLOGICALRESEARCH INCLUDINGTHENATIONALECOLOGICALOBSERVA TORYNETWORK.%/. REQUIREMENTFORCORE WILDLAND AREAS AS REMOTE MEASUREMENT SITES!CQUIRINGORNEGOTIATINGDEVELOPMENT RIGHTS FOR REMOTE AREAS ON PRIVATE LANDS AND INCREASING THE NUMBER OF DESIGNATED WILDERNESS AREAS ON PUBLIC LANDS FOR IN CORPORATIONINTOALARGERRESERVENETWORK WOULDFACILITATETHEPRESERVATIONOFREMAIN ING CORE HABITAT ASSOCIATED BIOLOGICAL DIVERSITY AND KEYSTONE SPECIES "ECAUSE MANY OF THE DATA SETS WE USED ARE AVAIL ABLENATIONWIDE SIMILARANALYSESCOULDBE CONDUCTEDTOASSESSTHEEXTENTANDSTATUS OFCOREAREASACROSSOTHERREGIONS 7HILEOURRESULTSHIGHLIGHTIMPORTANTISSUES FACINGCOREAREASINTHEEAST THEYUNDER SCOREIMPORTANTISSUESOFDATAQUALITYAND COMPLETENESSWHENCONDUCTINGLARGESCALE ASSESSMENTS&OREXAMPLE IMPROVEDROAD DATACOULDALTERTHERESULTSOFTHEANALYSIS PRESENTED HERE )F ROADS WERE OMITTED PRIMARILY AT PATCH EDGES MANY SMALLER PATCHES WOULD BE ELIMINATED LEADING TO HIGHER MEAN AND MEDIAN PATCH AREAS )F ROADS WERE OMITTED INTERNAL TO PATCHES ELIMINATIONOFSMALLAREASCOULDBEOFFSET BYSPLITTINGOFLARGERAREAS LEADINGTONO NETCHANGEOREVENDECREASESINMEANAND MEDIAN PATCH AREA 0ERIMETER AREA RATIO COULD DECREASE IF MANY SMALL PATCHES WEREELIMINATEDORINCREASEIFTHESPLITSOF LARGEPATCHESOFFSETSMALLPATCHLOSS7E WOULDEXPECTTHELARGESTEFFECTSIN'!0 ANDPRIVATELANDSWHERESMALLERROADSFOR TIMBEREXTRACTIONAREMORECOMMON ! MORE CURRENT DATABASE OF PROTECTED AREAS COULD ALSO ALTER OUR VIEW OF THE STATUS OF CORE AREAS -ANY .'/S IN THE %ASTERN5NITED3TATES NOTABLY4HE.ATURE #ONSERVANCY AND THE 7ILDLANDS 0ROJECT ARE CURRENTLY WORKING ON THE PROTECTION OF LARGE PARCELS HUNDREDS OF THOUSANDS OF HECTARES MUCH OF IT FORESTED VIA CONSERVATION EASEMENTS MANAGEMENT AGREEMENTS OR FEE SIMPLE ACQUISITIONS 4HUS WHILEOURRESULTSDONOTREmECTTHE CURRENT EXTENT OF PRIVATE LANDS PROTECTED FROMDEVELOPMENT THEYDOEMPHASIZETHE CONTRIBUTIONPRIVATELANDSCANMAKETOTHE CURRENTRESERVENETWORK

Volume 28 (3), 2008

ACKNOWLEDGMENTS 7E ACKNOWLEDGE SUPPORT FROM 7OODY 4URNER VIA THE .!3! !PPLIED 3CIENCES 0ROGRAM %COLOGICAL &ORECASTING AND WETHANK"EN"ESTFORHISASSISTANCEWITH THEHABITATMODELINGTOOLBOXHEDEVELOPED AT $UKE 5NIVERSITY 7E ALSO ACKNOWL EDGE THE ASSISTANCE OF 'REG &ISKE $AN 3TEINBERG AND #LAIRE *ANTZ WITH DATA SET PROCESSING 0ATRICK *ANTZ IS A DOCTORAL STUDENT AT THE "REN 3CHOOL OF %NVIRONMENTAL 3CI ENCE-ANAGEMENTATTHE5NIVERSITYOF #ALIFORNIA 3ANTA "ARBARA (IS RESEARCH FOCUSES ON THE INTERACTIONS BETWEEN HU MAN SETTLEMENT PATTERNS LAND COVER AND BIODIVERSITY 3COTT 'OETZ WORKS ON THE APPLICATION OF SATELLITE IMAGERY TO ANALYSES OF ENVIRON MENTALCHANGE INCLUDINGMONITORINGAND MODELINGLINKSBETWEENLANDUSECHANGE FOREST PRODUCTIVITY BIODIVERSITY CLIMATE AND HUMAN HEALTH "EFORE JOINING THE 7OODS(OLE2ESEARCH#ENTER HEWASONTHE FACULTYATTHE5NIVERSITYOF-ARYLANDFOR SEVENYEARS WHEREHEMAINTAINSANADJUNCT ASSOCIATEPROFESSORAPPOINTMENT ANDWAS A RESEARCH SCIENTIST AT .!3!S 'ODDARD 3PACE&LIGHT#ENTER(ERECEIVEDHIS0H$ FROMTHE5NIVERSITYOF-ARYLAND

LITERATURE CITED "ROWN $' +- *OHNSON 42 ,OVELAND AND $- 4HEOBALD  2URAL LAND USE TRENDS IN THE CONTERMINOUS 5NITED 3TATES  %COLOGICAL!PPLICATIONS   "UNN ! $, 5RBAN AND 4( +EITT  ,ANDSCAPE CONNECTIVITY A CONSERVATION APPLICATIONOFGRAPHTHEORY*OURNALOF%N VIRONMENTAL-ANAGEMENT  #ARROLL # 2% .OSS 0# 0AQUET AND .( 3CHUMAKER  %XTINCTION DEBT OF PROTECTED AREAS IN DEVELOPING LANDSCAPES #ONSERVATION"IOLOGY  $ELLA3ALA $! ., 3TAUS *2 3TRITTHOLT ! (ACKMAN AND ! )ACOBELLI  !N UPDATED PROTECTED AREAS DATABASE FOR THE 5NITED3TATESAND#ANADA Natural Areas Journal  $E6ELICE 2, AND *2 -ARTIN  !S

SESSING THE EXTENT TO WHICH ROADLESS AREAS COMPLEMENTTHECONSERVATIONOFBIOLOGICAL DIVERSITY%COLOGICAL!PPLICATIONS  %LVIDGE #$ #-ILESI *"$IETZ "44UTTLE 0#3UTTON 2.EMANI AND*%6OGELMANN  53 #ONSTRUCTED AREA APPROACHES THE SIZE OF /HIO !'5 %OS 4RANSACTIONS   %32) $ATA  -APS ;#$ 2/-=  53 7ATER "ODIES %NVIRONMENTAL 3YSTEMS 2ESEARCH)NSTITUTE )NC/RIGINn53'EO LOGICAL 3URVEY IN COOPERATION WITH 53 %NVIRONMENTAL0ROTECTION!GENCY &AHRIG , *(0EDLAR 3%0OPE 0$4AYLOR AND*&7EGNER%FFECTOFROADTRAFlC ONAMPHIBIANDENSITY"IOLOGICAL#ONSERVA TION  &ORMAN 244 AND ,% !LEXANDER  2OADS AND THEIR MAJOR ECOLOGICAL EFFECTS !NNUAL2EVIEWOF%COLOGYAND3YSTEMATICS   &ULLER 4+ 7% "ERG ', 2ADDE -3 ,ENARZ AND'"*OSELYN!HISTORY AND CURRENT ESTIMATE OF WOLF DISTRIBUTION ANDNUMBERSIN-INNESOTA7ILDLIFE3OCIETY "ULLETIN  'IBBS *0 AND 7' 3HRIVER  %STI MATING THE EFFECTS OF ROAD MORTALITY ON TURTLE POPULATIONS #ONSERVATION "IOLOGY   'LENNON -* AND7&0ORTER%FFECTSOF LANDUSEMANAGEMENTONBIOTICINTEGRITYAN INVESTIGATIONOFBIRDCOMMUNITIES"IOLOGI CAL#ONSERVATION  'ROSSMAN $( $ &ABER ,ANGENDOEN !3 7EAKLEY - !NDERSON 0 "OURGERON 2 #RAWFORD +'OODIN 3,ANDAAL +-ET ZLER +$0ATTERSON -0YNE -2EID AND ,3NEDDON)NTERNATIONAL#LASSIlCA TIONOF%COLOGICAL#OMMUNITIES4ERRESTRIAL 6EGETATIONOFTHE5NITED3TATES6OL) 4HE .ATIONAL6EGETATION#LASSIlCATION3YSTEM $EVELOPMENT 3TATUS AND!PPLICATIONS4HE .ATURE#ONSERVANCY !RLINGTON 6A 'OETZ 3* 2EMOTESENSINGOFRIPARIAN BUFFERSPASTPROGRESSANDFUTUREPROSPECTS Journal oF the American Water Resources Association   'OETZ 3* AND0*ANTZ3ATELLITEMAPS SHOW#HESAPEAKE"AYURBANDEVELOPMENT !'5 Eos Transactions  (AGAN *- ,# )RLAND AND!!7HITMAN  #HANGING TIMBERLAND OWNERSHIP IN THE .ORTHERN &OREST AND IMPLICATIONS FOR BIODIVERSITY 2EPORT -##3 �   -ANOMET#ENTERFOR#ONSERVATION3CIENCES "RUNSWICK -AINE (ANNAH , '& -IDGLEY 4 ,OVEJOY 7*

Natural Areas Journal 273

"OND -"USH *#,OVETT $3COTT AND &)7OODWARD#ONSERVATIONOFBIODI VERSITYINACHANGINGCLIMATE#ONSERVATION "IOLOGY  (ANSEN !* 2, +NIGHT *- -ARZLUFF 3 0OWELL +"ROWN 0('UDE AND+*ONES  %FFECTS OF EXURBAN DEVELOPMENT ON BIODIVERSITY PATTERNS MECHANISMS AND RESEARCH NEEDS %COLOGICAL !PPLICATIONS   ( A N S E N -  #  2  3  $ E & R I E S *  2  '  4OWNSHEND - #ARROLL # $IMICELI AND 2! 3OHLBERG  'LOBAL PERCENT TREE COVERATASPATIALRESOLUTIONOFMETERS lRSTRESULTSOFTHE-/$)3VEGETATIONCON TINUOUSlELDSALGORITHM%ARTH)NTERACTIONS   (ANSEN -# 23$E&RIES *2'4OWNSHEND ,-ARUFU AND23OHLBERG$EVELOP MENTOFA-/$)3TREECOVERVALIDATIONDATA SETFOR7ESTERN0ROVINCE :AMBIA2EMOTE 3ENSINGOF%NVIRONMENT  (AWBAKER 4* AND 6# 2ADELOFF  2OADS AND LANDSCAPE PATTERN IN .ORTHERN 7ISCONSIN BASED ON A COMPARISON OF FOUR ROAD DATA SOURCES #ONSERVATION "IOLOGY   )RWIN %' AND .% "OCKSTAEL  )NTER ACTING AGENTS SPATIAL EXTERNALITIES AND THE EVOLUTION OF RESIDENTIAL LAND USE PATTERNS *OURNALOF%CONOMIC'EOGRAPHY  *ANTZ 0 3'OETZ AND#*ANTZ5RBAN IZATIONANDTHELOSSOFRESOURCELANDSINTHE #HESAPEAKE"AYWATERSHED%NVIRONMENTAL -ANAGEMENT  ,OUCKS # . "ROWN ! ,OUCKS AND + #ESAREO53$!FORESTSERVICEROAD LESSAREASPOTENTIALBIODIVERSITYCONSERVA TION RESERVES #ONSERVATION %COLOGY  !VAILABLE ONLINE HTTPWWWCONSECOL ORGVOLISSART -ADER (*  !NIMAL HABITAT ISOLATION BYROADSANDAGRICULTURALlELDS"IOLOGICAL #ONSERVATION  -C'ARIGAL + 3!#USHMAN -#.EEL AND %%NE&2!'34!433PATIAL0ATTERN

274 Natural Areas Journal

!NALYSIS 0ROGRAM FOR #ATEGORICAL -APS #OMPUTERSOFTWAREPROGRAMPRODUCEDBYTHE AUTHORSATTHE5NIVERSITYOF-ASSACHUSETTS !MHERST !VAILABLE ONLINE WWWUMASS EDULANDECORESEARCHFRAGSTATSFRAGSTATS HTML -C'RATH $! *0%VANS #+3MITH $' (ASKELL .70ELKEY 22'OTTFRIED #$ "ROCKETT -$,ANE -$ AND%$7IL LIAMS-APPINGLAND USECHANGEAND MONITORINGTHEIMPACTSOFHARDWOOD TO PINE CONVERSION ON THE SOUTHERN #UMBERLAND 0LATEAU IN 4ENNESSEE %ARTH )NTERACTIONS   -ECH ,$  7OLF POPULATION SURVIVAL IN AN AREA OF HIGH ROAD DENSITY!MERICAN -IDLAND.ATURALIST  -ECH ,$ 3(&RITTS AND$7AGNER -INNESOTAWOLFDISPERSALTO7ISCONSINAND -ICHIGAN !MERICAN -IDLAND .ATURALIST   -LADENOFF $* 4!3ICKLEY 2'(AIGHT AND !07YDEVEN!REGIONALLANDSCAPE ANALYSIS AND PREDICTION OF FAVORABLE GRAY WOLF HABITAT IN THE NORTHERN 'REAT ,AKES REGION#ONSERVATION"IOLOGY  .OSS 2& *23TRITTHOLT +6ANCE "ORLAND # #ARROLL AND0&ROST!CONSERVATION PLAN FOR THE +LAMATH 3ISKIYOU ECOREGION .ATURAL!REAS*OURNAL  /"RIEN $ --ANSEAU !&ALL AND-*&OR TIN4ESTINGTHEIMPORTANCEOFSPATIAL CONlGURATIONOFWINTERHABITATFORWOODLAND CARIBOU AN APPLICATION OF GRAPH THEORY "IOLOGICAL#ONSERVATION  0ARENDES ,! AND *! *ONES  2OLE OFLIGHTAVAILABILITYANDDISPERSALINEXOTIC PLANTINVASIONALONGROADSANDSTREAMSINTHE (*!NDREWS%XPERIMENTAL&OREST /REGON #ONSERVATION"IOLOGY  2ADELOFF 6# 2"(AMMER 3)3TEWART *3 &RIED AND*&-C+EEFRY4HEWILD LAND URBAN INTERFACE IN THE 5NITED 3TATES %COLOGICAL!PPLICATIONS  3COTT *- &$AVIS "#SUTI 2.OSS ""UT TERlELD #'ROVES (!NDERSON 3#AICCO

&$ERCHIA 4%DWARDS *5LLIMAN AND2' 7RIGHT'APANALYSISAGEOGRAPHICAP PROACHTOPROTECTIONOFBIOLOGICALDIVERSITY 7ILDLIFE-ONOGRAPHS  3HEPHERD " AND * 7HITTINGTON  2E SPONSE OF WOLVES TO CORRIDOR RESTORATION ANDHUMANUSEMANAGEMENt. %COLOGYAND 3OCIETY!VAILABLEONLINEHTTPWWW ECOLOGYANDSOCIETYORGVOLISSART 3TRITTHOLT *2 AND$!$ELLASALA)M PORTANCE OF ROADLESS AREAS IN BIODIVERSITY CONSERVATION IN FORESTED ECOSYSTEMS CASE STUDY OF THE +LAMATH 3ISKIYOU ECOREGION OFTHE5NITED3TATES#ONSERVATION"IOLOGY   4HEOBALD $-  ,AND USE DYNAMICS BEYOND THE !MERICAN URBAN FRINGE 'EO GRAPHICAL2EVIEW  4HIEL 20  4HE RELATIONSHIP BETWEEN ROAD DENSITIES AND WOLF HABITAT IN 7IS CONSIN !MERICAN -IDLAND .ATURALIST   4ILMAN $ 2--AY #,,EHMAN AND-! .OWAK  (ABITAT DESTRUCTION AND THE EXTINCTIONDEBT.ATURE  4URNER *- #ONSERVATIONSCIENCEAND FOREST SERVICE POLICY FOR ROADLESS AREAS #ONSERVATION"IOLOGY  5RBAN $,-ODELINGECOLOGICALPROCESS ESACROSSSCALES%COLOGY  5RBAN $, AND 4 +EITT  ,ANDSCAPE CONNECTIVITYAGRAPH THEORETICPERSPECTIVE %COLOGY  53&OREST3ERVICE2OADLESSAREACON SERVATIONlNALENVIRONMENTALIMPACTSTATE MENT APPENDIX!INVENTORIEDROADLESSAREA ACREAGECATEGORIESOF.&3LANDSSUMMARIZED BYREGION53$EPARTMENTOF!GRICULTURE 7ASHINGTON $#!VAILABLEONLINEHTTP ROADLESSFSFEDUSDOCUMENTSFEIS 7ATTS 2$ 27 #OMPTON *( -C#AM MON #, 2ICH 3- 7RIGHT 4 /WENS 4 AND$3/UREN2OADLESS3PACE OFTHE#ONTERMINOUS5NITED3TATES3CIENCE  

Volume 28 (3), 2008