Lynx habitat use - cpw.state.co.us...of habitat was below 0.102 units per ha (~1 unit per 10 ha)....
19
1 Areas of high habitat use from 1999‐2010 for radio‐collared Canada lynx reintroduced to Colorado David M. Theobald, PhD. Department of Fish, Wildlife, and Conservation Biology Colorado State University Fort Collins, CO 80523‐1474 and Tanya M. Shenk* Colorado Division of Wildlife, Mammals Research Fort Collins, CO 80526 31 March 2011 *current affiliation National Park Service, Biological Resources Management Division, Fort Collins, CO 80525 Introduction The purpose of this report is to describe an analysis of current habitat use for the 218 Canada lynx that were reintroduced to Colorado from 1999 to 2006. The primary dataset used here is location data from collared individuals, and collected by the Colorado Division of Wildlife from 1999 until present. Data from individual animals were combined together to form a “population‐level” estimate of habitat use by weighting locations based on the number of months data were collected for an individual. Basic descriptive and summary statistics such as a cumulative distribution function provide relative proportion of use in a given class of habitat. Note that this study was not intended to examine individual home range size, territoriality, or movement relative to land use and/or transportation corridors, nor is it intended to predict potential or future habitat use. Methods Preparation of location dataset I received two datasets on lynx locations from CDOW (Tanya Shenk and Jake Ivan, personal communication) dated November 9, 2010. The VHF dataset was collected by locating individual lynx via telemetry during fixed‐wing airplane flights. Most of the locations were south of I‐70, as monitoring was focused on observing animals in the core release area, roughly defined as the high elevation areas in southwestern Colorado bounded by Taylor Mesa on the west, Gunnison basin on the north, Poncha Pass on the east, and New Mexico border on the south. Aerial locations were obtained outside of the core area on an opportunistic basis, typically only 1 location per 3 months. The entire VHF dataset had 11,356 observations for 257 individuals (103 females, 117 males) collected from 2/4/1999 to 6/22/2010. The Argos data were collected from lynx that were outfitted with dual VHF/Argos satellite collars, beginning in April 2000. These collars were designed to provide locations
Transcript of Lynx habitat use - cpw.state.co.us...of habitat was below 0.102 units per ha (~1 unit per 10 ha)....
1
Areasofhighhabitatusefrom1999‐2010forradio‐collaredCanadalynxreintroducedtoColorado
DavidM.Theobald,PhD.DepartmentofFish,Wildlife,andConservationBiology
ColoradoStateUniversityFortCollins,CO80523‐1474
andTanyaM.Shenk*
ColoradoDivisionofWildlife,MammalsResearchFortCollins,CO80526
31March2011*currentaffiliationNationalParkService,BiologicalResourcesManagementDivision,FortCollins,CO80525
IntroductionThepurposeofthisreportistodescribeananalysisofcurrenthabitatuseforthe
218CanadalynxthatwerereintroducedtoColoradofrom1999to2006.Theprimarydatasetusedhereislocationdatafromcollaredindividuals,andcollectedbytheColoradoDivisionofWildlifefrom1999untilpresent.Datafromindividualanimalswerecombinedtogethertoforma“population‐level”estimateofhabitatusebyweightinglocationsbasedonthenumberofmonthsdatawerecollectedforanindividual.Basicdescriptiveandsummarystatisticssuchasacumulativedistributionfunctionproviderelativeproportionofuseinagivenclassofhabitat.Notethatthisstudywasnotintendedtoexamineindividualhomerangesize,territoriality,ormovementrelativetolanduseand/ortransportationcorridors,norisitintendedtopredictpotentialorfuturehabitatuse.
Methods
PreparationoflocationdatasetIreceivedtwodatasetsonlynxlocationsfromCDOW(TanyaShenkandJakeIvan,
personalcommunication)datedNovember9,2010.TheVHFdatasetwascollectedbylocatingindividuallynxviatelemetryduringfixed‐wingairplaneflights.MostofthelocationsweresouthofI‐70,asmonitoringwasfocusedonobservinganimalsinthecorereleasearea,roughlydefinedasthehighelevationareasinsouthwesternColoradoboundedbyTaylorMesaonthewest,Gunnisonbasinonthenorth,PonchaPassontheeast,andNewMexicoborderonthesouth.Aeriallocationswereobtainedoutsideofthecoreareaonanopportunisticbasis,typicallyonly1locationper3months.TheentireVHFdatasethad11,356observationsfor257individuals(103females,117males)collectedfrom2/4/1999to6/22/2010.
TheArgosdatawerecollectedfromlynxthatwereoutfittedwithdualVHF/Argossatellitecollars,beginninginApril2000.Thesecollarsweredesignedtoprovidelocations
2
onceperweek.TheArgosdatasethad33,778observationsfor196individuals(97females,88males),collectedfrom3/1/2000to8/11/2010.
Basedonanumberofdiscussionswithlynxbiologistsandstatisticians1,wefilteredthedatasetsinthefollowingways:
a. Removethefirst6monthsoflocationsafterreleaseforeachindividualtoreducelikelybiasof“justreleased”movements.Thisresultedin35,276locationsdatingfrom9/13/1999to8/11/2010,representing198individuals(152with>30locations).
b. RemoveArgoslocationswithhighspatialuncertainty(Table1;LocationClass1,0,A,B,Z)andretaintherest(LocationClass3and2).Thisresultedin15,545locationsrepresenting197individuals(129with>30locations).
c. Removelocationsthatrepresentmultiplefixesinaday,retainingthemostpreciselocationestimate(VHF,Argos3,Argos2).Thisresultedin13,803locationsrepresenting197individuals(118with>30locations).
d. Removerecordsforindividualswithlessthan30locations,resultingin12,796locationsfrom118individuals
Thus,thefinaldatasetusedintheanalysisincluded12,796observationsfor118individuals(Figures1&2;numberofindividuals:f=64,m=54;totalmonthsofdata:f=2679,m=1784).NotethatIdidnotconductananalysisthatseparatessummerfromwinteruse,ormalesfromfemales.Figure3showsthedistributionoflynxlocationsbyyear.Table1.Spatialuncertaintyofthelocationdata.*indicatesdatatypesusedintheanalysis.
Locationsource/class
Description Radius(m)encompassing68%oferrordistribution
Radius(m)encompassing95%oferrordistribution
*VHF Datacollectedbytelemetry 200 400*Argos3 <250m 250 500*Argos2 250‐500m 375 750Argos1 1500m 1500 3000Argos0 N/A Notused ArgosA,B,Z N/A Notused
1WithTanyaShenk(NPS)andJakeIvan,PaulLukacs,andMindyRice(CDOW)
3
Figure1.Thenumberofobservationsandmonthsfor118individuallynxusedintheanalysisofhabitatuse.
Figure2.Thedistributionof118individuallynxusedintheanalysisofhabitatuse.Differentcolorsdenotedifferentindividuals.
4
Figure3.Thedistributionoflynxlocationsusedinthisanalysis,displayedbyyearfrom1999to2010.
5
6
7
GenerateutilizationdistributionsIfollowedthegeneralapproachofMillspaughetal.(2006)topreparetheutilization
distributionsurfaceandexplanatoryvariables.First,Iestimatedtheutilizationdistribution(UD)foreachindividualanimalusinga
homerangeestimatorcalledlocalconvexhulls(LoCoH;GetzandWilmers2004).Weselectedthisapproachbecauseitisnon‐parametric,producesaUD,andidentifiesabrupt“edges”inthespatialdistributionthatcanoccurbecauseofedgesinlandscapefeatures(e.g.,topographicconstraints)orterritorialityamongindividuals.Generally,LoCoHextendstheconceptofconvexhullstodelineatespaceuse,butratherthanencompassingallpointsatonce(e.g.,minimumconvexpolygon),itworksonasub‐setofpointstoidentifylocalconvexhulls.Thatis,asmallor“local”convexhullisidentifiedaroundpointi,whichcontainsasetofknearestneighbors.Thelocalhullsarethensortedbyarea,smallesttolargest,andtheUDvalue(orisopleths)aredeterminedbytheproportionofpointsfoundineachlocalhull.
Specifically,weusedtheadaptiveversioncalleda‐LoCoHbecauseitismostinsensitivetosub‐optimalvalueparameterization(Getzetal.2007).a‐LoCoHidentifiesavariablenumberofknearestneighborssuchthatitusesallpointswithinavariablecirclearoundarootpointsothatthesumofdistancebetweenthepointsandtherootpointdoesnotexceedauser‐definedthresholdvalue,a.Thismethodadjuststheradiussothatsmallerconvexhullsariseinhighuseareasandprovidemoreclearlydefinedisoplethsinregionswithhigherdensityoflocationdata.Getzetal.(2007)recommendedsettingatothemaximumdistancebetweenlocationpoints(withaminimumofk=3),whennootheraprioriinformationisavailable.Iusedtheaverageofthewidthandheightofthemaximumenclosingconvexhullbecausehabitatuseoftenoccurredindisjunctclusters.ToremoveartifactsintheLoCoHoutputthatcanbeintroducedbyspuriouslocationvalues,Iremovedthe100%isopleth(leavingallisopleths<=90%).EachUDwasnormalizedsothattheareaundertheUDsummedto1.0.
Second,tocombinetheUDsforthe118individualsintoageneral,population‐levelestimateofhabitatuse,Icomputedthenumberofmonthsoflocationdataforeachindividual.Thisprovidedaweightsuchthatindividualswhohadarelativelyshortdurationoflocationswouldhaveminorinfluence,whileanindividualwithalongduration(manymonths)oflocationdataavailablewouldhavemoreinfluenceonthepopulation‐levelUDsurface.Theweightsrangedfrom11to113,withameanof37.82(SD=18.49).Foreachcellwithinanindividual,normalizedUDvaluesweremultipliedbytheappropriateweightandtheresultwassummedateachrastercellacrossallindividuals.2Thus,ifonecellorgeneralareaisusedintenselybyonlyoneindividual,itwillbehighintensityuse.Conversely,ifnumerousindividualshaveusedalocation,eachatalowerintensity,ittoo,willhaveahighUD.
Third,IsmoothedtheUDvaluesfromaresolutionof90mtoamorebiologically‐basedresolutionthatwasequaltotheareaofthesmallestobserved“highuse”UDobservedacrossallindividuals–225.0ha.Iaggregatedthe90mcellstoacellsizeof1440m(~207ha).3Thisprovides11,934cells(at1,440m)inthepopulation‐levelUD(pUD).2Multipliedby100,000,000anddividedbytotaltogetintegergrid,from1‐1701.3Aggregatedbasedonthesmallesthigh‐qualityhabitatfrom90to1440musingMEAN,notamovingwindowaverage.
8
Generatehabitatvariables BasedonpreviousworkwithCDOWlynxbiologiststhatidentifiedlikelyimportantexplanatoryvariablesforhabitatuse,Idevelopedasetoflandscape‐leveldatasetsbasedonregionally‐availabledatasets(Table2).ThesedataincludedcompositionalmeasuresoftheproportionofvariousvegetationtypescomputedfromLANDFIREexistingvegetationtypeclasses(30m),topographicvariablescalculatedfromtheUSGSNationalElevationDataset(30m),anddistancevariablescomputedfromColoradoDepartmentofTransportationhighwaysdata.Inaddition,Iexaminedtherelationshipofhabitatusetoindividualtreespeciestodiscernwhatspecificforestedvegetationstypeswereusedmore(orless)bylynx.Table2.Descriptionofthelandscape‐levelvariablesusedtoexaminehabitatuseoflynx.
Type Variable DescriptionBHD Censusblockhousingdensity(SERGoMv1.1,units*1000per
ha).DatasourcefromTheobald(2005),Bierwagenetal.(2010)Composition
RDENS Roaddensity(km/km2)Lf1 Water(Lakes,reservoirs,largerivers)LF2 Rock,snow,iceLF3 Urban/built‐upLF4 Agriculture(cropland,pasture)LF5 Forest(uppermontane)–Spruce‐fir,subalpine,lodgepole,
mixedaspen‐conifer,DouglasfirLF6 Forest(lowermontane)–ponderosapine,pinyon‐juniperLF7 ShrublandsLF8 Grasslands–includessub‐alpinemeadowsandalpinetundraLF9 Shrub(steppe)
Proportion
LF10 Riparian&wetlandsDEM Elevation(meters)SLOPE Slope–averageslopeindegrees(computedfrom30m)
Topographic
TWIP TopographicwetnessindexplussolarinsolationD4P5HA Forest(mesic)patchesatleast50haD4r2k HighwayswithAADT<2kD4R2_5K Highwayswith2k<=AADT<5kD4R5_10K Highwayswith5k<=AADT<10kDR410K HighwayswithAADT>10k
Distance
D4RDS Allnon‐highwayroads
9
ResultsFollowingconvention,Idefinedlynxhabitatasareaswithinthe90%isopleth.Over
3.5millionacresinColoradoandNewMexicowerefoundtobecurrentlywithinlynxhabitat.Figure4showsthedistributionofthepopulation‐levelutilizationdistributionsurface,whileFigure5showsthenumberofindividualsthatwerefoundineachoftheUDpolygonsandFigure6depictsacumulativedistributionfunctionoftheUDvalues.Table3providesasummaryofthelandscapevariablesforlynxhabitat.
Theaverageelevationforlynxhabitatwas3,285m(10,780ft),withthemajority(68%or+and–1SD)ofhabitatlocatedbetween3,027and3,543m(9,900‐11,620ft).Theaverageslopeforlynxhabitatwas18.9(withthemajoritybetween12.8and25.1degrees).Theaveragetopographicwetnessindexvalue(TWI+)was3.38,rangingfrom2.64to4.12(lowvaluesindicatehighsoilmoisturenearthefootofslopesonnorthaspectswhilehighvaluesindicatelowsoilmoistureindicativeofridgetopsandsouth‐facingaspects).
Housingdensityinlynxhabitatwaslow,withameanvalueof0.011unitsperha(~1unitper80ha);themajorityofhabitatwasbelow0.102unitsperha(~1unitper10ha).Roaddensityinlynxhabitatwasalsolow,withameanvalueof0.51km/km2,andthemajorityofhabitatbelow1.22km/km2.
Theaverageproportionofforest(uppermontane)inlynxhabitatwas0.65,withthemajorityoccurringinareaswithatleast20%forested(uppermontane)cover.Habitatusewasalsoassociatedwithdistancefromlargepatches(>50ha)offorest(uppermontane)cover,withthemajorityofhabitatwithin3.35km,andtheaverageat0.36km.Theaverageproportionofgrasslandswas0.16.Therewaslittleassociationoflynxhabitatuseareaswithotherlandcovertypes. Lynxhabitatuseareasoccurredawayfromhighwayswithhightraffic(AADT>10k),averagingatleast43km,withmajorityatleast27kmaway.Thisdeclinedtobetween25.9and16.0kmaveragedistanceforotherhighwaytypes,withthemajorityofhabitatbeingatleast5.2kmfromthenearesthighway.Table4providesasummaryofthemostcommonlandcovertypes(fromLANDFIRE)foundtooccurinlynxhabitatuseareas.Appendix4providesscatterplotsforspeciesspecificforestedvegetationtypes.Subalpine/spruce‐firforestdominatestheUDpolygonsareawith43.3%.Otheruppermontaneandtundracovertypesareidentified,combiningtoover86%.Becauseofthegrainofthevegetationdata(aggregatedto2.25km2),covertypessuchasbarren/rockandtundracoverlikelyincludesmallstandsofforestandripariancover.
10
Table3.Summaryoflandscapevariablesforthelynxhabitatuseareas.Unitsareinproportion(0.01.0)ifnototherwisedenoted.AlsoseeAppendix1forcumulativedistributionsofthesevariablesandAppendix2fordistributionmapsofeachvariable.
Type Variable Min -1 SD Mean +1 SD Max Housing density (units per ha) 0 N/A 0.011 0.102 10.531 Composition Road density (km/km2) 0 N/A 0.513 1.22 12.7Water (Lakes, reservoirs, large rivers) 0 N/A 0.005 0.03 0.9Rock, snow, ice 0 N/A 0.063 0.16 0.9Urban/built-up 0 N/A 0.002 0.01 0.7Agriculture (cropland, pasture) 0 N/A 0.003 0.04 0.95Forest (upper montane) 0 0.4201 0.653 0.89 0.99Forest (lower montane) 0 N/A 0.009 0.06 1Shrublands 0 N/A 0.008 0.06 0.95Grasslands 0 N/A 0.163 0.36 1Shrub (steppe) 0 N/A 0.061 0.14 0.94
Proportion (01)
Riparian & wetlands 0 N/A 0.031 0.06 0.56Elevation (meters) 1399 3027 3285 3543 4143 Slope (degrees) 0.1 12.8 18.9 25.1 37.6
Topographic
Topographic wetness index plus 1.40 2.64 3.38 4.12 15.20Highways with AADT ≥10k 0.20 27.86 43.88 59.91 77.00 Highways with 5k<=AADT<10k 0.20 14.59 25.96 37.32 51.80Highways with 2k<=AADT<5k 0.20 6.15 16.05 25.95 39.90Highways with AADT <2k 0.20 5.22 14.35 23.48 40.70All non-highway roads 0.03 N/A 1.88 4.26 35.00
Distance (km)
Forest (upper montane) patches ≥50 ha 0.00 N/A 0.36 3.35 27.10
11
Figure4.Thepopulation‐levelutilizationdistributionfor118lynxintheanalysisdataset,shownwithmajorhighwaysandcountyboundariesforreference.Low‐intensityuseisshowninyellow,moderateinorange,highinblue.
12
Figure5.Thenumberofindividuallynxthatwerefoundinpolygonsofthepopulation‐levelutilizationdistribution,shownwithmajorhighwaysforreference.
13
Figure6.Thecumulativedistributionfunctionforthepopulation‐levelutilizationdistribution.Roughly50%oftheUDhasvalueslessthan12,about25%between12and30,andthetop25%between31and950.
Table4.Theproportionofdominantlandcovertypes(fromLANDFIRE)occurringinlynxhabitatuseareas.(Allremainingcovertypesarelessthan1%).
Existingvegetationtype PercentageSubalpine/Spruce‐firforest 43.3Barren/rock/tundra 13.3Alpinetundra(“turf”) 7.8Aspenforest 7.1Aspen‐mixedconiferforest 6.8Subalpine/uppermontaneriparian 6.3Snow‐ice 2.1Subalpine/spruce‐fir(mesic) 1.7Subalpinemontanemeadow 1.5
14
Discussion/conclusionThisreportprovidesthefirstestimateoftheoverall,population‐levelhabitat
currentlyusedbylynxinColorado,withover3.5millionacres.ThemajorityofthecurrentlynxuseareasarelocatedonUSForestServicelands(Figure7).TwolargecontiguousareasofhabitatusearefoundintheSanJuanmountainrangeandtheCollegiatePeaksrangingnorthofMonarchPasstoVailPassandspanningI‐70nearLovelandPasstotheFrasierExperimentalForest.Threeothersmallerareaswereidentified,ontheGrandMesa,intheWestElksjustnorthofBlackCanyonoftheGunnison,andanareacenteredaroundRockyMountainNationalPark.Elevenof25downhillskiareasinColoradoarelocatedinlynxhabitat.
ThecentralfindingsofthisanalysisareconsistentwithpreviousreportsthattheCanadalynxreintroducedtoColoradohaveprimarilyusedhighelevationspruce‐firandaspenvegetationtypesashabitat.Thesereportsarebasedonvegetationdatacollectedduringaerialsurveysaswellas“ontheground”snow‐tracksurveys(CDOW2009;Shenk2009).
15
Figure7.–TheutilizationdistributionforcurrentlynxhabitatinColorado,withforestserviceadministrativeboundaries.
16
Highuseareas(thosewithlargerUDvalues)arecharacterizedbyahighpercentageofuppermontaneforest,highelevations,andhighmoisturepositions(i.e.lowTWI+values;Figure8andAppendix4).Withinforestvegetationtypes,thestrongestrelationshiptoUDweresub‐alpine/Spruce‐fir,andaspenandaspen/mixedconifervegetationtypes(Figure9).Figure8.ComparisonofUDforcurrentlynxhabitatversusuppermontaneforest(upperleft),Topographicwetnessindexplus(upperright),andelevation(lowerleft).
17
Figure9.Forestvegetationtypeswithstrongpositiveassociationwithcurrentlynxhabitat:subalpine/spruce‐fir(left)andaspen/aspen‐mixedconifer(right).
LimitationsofanalysisAkeylimitationtothefindingshereisthatthedatarepresentonlyasampleoflynxin
thestate.Initially,100%oftheindividuallynxwerecollared,butthepercentoflynxinthestatethatweretrackeddiminishedovertime.Forexample,thenumberoflynxsampledin2010representsabout¼ofthenumbersampledduring2005(Table5).
Table5.Annualsummaryofnumberofindividualsandlocations.Year #individuals #locations
1999 9 782000 40 4432001 38 7982002 35 10992003 51 10112004 70 18112005 95 20822006 85 18912007 73 15462008 57 9362009 43 7862010 27 315
Wheninterpretingtheresults,itisimportanttorecallthattheanalysisisrestrictedto
areaswithlynxhabitatuse.Thatis,even“low”useareasprovideimportanthabitat,itisjustrelativelylowercomparedtohighuseareas.Also,rememberthatwhenthinkingabouttherelationshipbetweenuseandthevariouslandscapevariables,thevariouscovertypesarerestrictedtothehabitatuseareas,notthefullextentofcovertypeswithinabroader
18
area(e.g.,corereleasesite,southernrockiesecoregion,etc.).AreasthatareoutsideoftheUDpolygonsidentifiedinthisanalysismaystillbeused,andtheUDpolygonsofhabitatusemightchangewithfuturedistributionsoflynx.
Becarefultointerpretthelandscapevariablesthatareestimatedoriginallyat30mresolution(fromLANDFIRE)becausemanyfine‐grainvariables,suchasnarrowriparianareasarenotrecordedinthesedata.Also,thesedatadonotcapturevariationinunder‐storyvegetation,ratherthesatelliteimagerypredominatelycapturesover‐storyconditions.
RecommendationsforfurtheranalysisTherearethreelogicalnextstepsfortheanalysisoflynxdistributionsinColorado:(a)comparisonoftheseresultstofield‐collectedvegetationdataonunderstoryandtopographiccharacteristics;(b)predictionofsuitablelynxhabitatbeyondthecorereintroductionarea;and(c)examinationofmovementincludinghighwaycrossinglocations.
a. Afirststepistocomparethefindingsofthisreportbasedonlynxlocationsandlandscape‐levelvariablesagainstthesite‐scalehabitatdatacollectedbyCDOWduringsnow‐trackingoflynx(forwinterseasonsforallyears).Ateachsite,datawererecordedonthelynxtracked,slope,aspect,foreststructureclass(grass/forb,shrub/seedling,sapling/pole,mature,andoldgrowth),location,andelevation(Shenk2006).Thesedatahavebeencompiledintoadatabaseandsummarizedinreports(e.g.,Shenk2009).Currently,CDOWareconductingaqualitycontrolprocesstoenablemoredetailed,spatially‐explicitcomparisonofthesite‐levelhabitatvariablesagainstlandscape‐levelvariables(JakeIvan,personalcommunication).
b. Asecondstepistodevelopaspatially‐explicitmodelthatpredictshabitatusewithinthestateofColorado,basedontherelationshipsbetweenlynxlocationsandlandscapevariablessuchasthosedescribedinthisreport.Generally,thetypeofmodelthatisappropriatedependsonthespecificmanagementquestionbeingasked,aswellasthenatureandqualityofthedatabeingused.Inparticular,understandingwhetherthelocationdataarebestconsideredpresence‐absencedataorpresence‐only.“…ifpresence‐absencesurveydataareavailable,webelieveitisgenerallyadvisabletouseapresence‐absencemodellingmethod,sinceinthatcasethemodelsarelesssusceptibletoproblemsofsampleselectionbias,thesurveymethodwilloftenbeknownandcanbeusedtoappropriatelydefinetheresponsevariableformodelling,andwetakeadvantageofallinformationinthedata(Elithetal.2011;pgs45‐46).Thelynxdatasetusedinthisreportandthatwouldbeavailableforaspatialpredictivemodelhavesomesamplingbiasissues,butshouldnotbeconsidered“presence‐only”data–becauseofthesystematiccollectionoflocationsviaVHF/aerialmethodsandthecomplementarydatafromARGOS.Certainlytherearesomeissueswiththelocationdata–suchasafocusondatacollectedsouthofI‐70andlikelymissingARGOSlocationsinhightopographicrelief/densevegetation–butcomparedwithmostotherwildlifestudiesofhabitatusethedatabaseisrobustgiventhelargenumberofindividualsanddurationofthestudy.TheCDOWhasheldpreliminarydiscussionsinDecember/January2011aboutpossiblemodelingapproachesandcandidatevariables.Aparticularchallengeistounderstandthe
19
potentialhabitatuseoflodgepolepine–sinceitissoprevalentinforestsnorthofI‐70,yetmuchoftheexisting(1999‐2010)locationsaredominatedbylocationsneartheSanJuancoreareathathasverylittlelodgepolepine.Dataonforeststandage/structureareapotentialsurrogateforunderstoryvegetationtypesthatarethoughttobeimportantforhigh‐qualityhabitat(forsnowshoehares).
c. Athirdstepistoexaminelynxmovementinrelationtotransportationandotherpotentialconflictinglanduses.Giventhespatialandtemporalresolutionofthelynxlocations,itisreasonabletoexaminethedatasetforsomebroad,landscape‐levelquantificationofhighwaycrossings.But,thedataarenotsuitedtoconductanalysestoidentifyspecificlocations(<~10km)wherelynxmaybecrossinghighwaysandtoexaminehowotherlandusesmightbeinfluencinghabitatquality.
LiteraturecitedBierwagen, B., D.M. Theobald, C.R. Pyke, A. Choate, P. Groth, J.V. Thomas, and P. Morefield.2010.Nationalhousingandimpervioussurfacescenariosforintegratedclimateimpactassessments.ProceedingsoftheNationalAcademyofSciences107(49):20887‐20892.ColoradoDivisionofWildlife(CDOW).2009.Lynxupdate.ReportpublishedMay25,2009.
Elith,J.,S.J.Phillips,T.Hastie,M.Dudik,Y.E.Chee,andC.J.Yates.2011.AstatisticalexplanationofMaxEntforecologists.DiversityandDistributions17:43‐57.Getz,W.M.andC.C.Wilmers.2004.Alocalnearest‐neighborconvex‐hullconstructionofhomerangesandutilizationdistributions.Ecography27:489‐505.
Getz,W.M.,S.Fortmann‐Roe,P.C.Cross,A.J.Lyons,S.J.Ryan,andC.C.Wilmers.2007.LoCoH:Nonparameterickernelmethodsforconstructinghomerangesandutilizationdistributions.PLoSONE2(2):e207.Millspaugh,J.J.,R.M.Nielson,L.McDonald,J.M.Marzluff,R.A.Gitzen,C.D.Rittenhouse,M.W.Hubbard,andS.L.Sheriff.2006.Analysisofresourceselectionusingutilizationdistributions.JournalofWildlifeManagement70(2):384‐395.Shenk,T.M.2006.WildlifeResearchReport20052006.ColoradoDivisionofWildlife,45pgs.
Shenk,T.M.2009.Lynxannualreport20082009.ColoradoDivisionofWildlife,55pgs.Theobald,D.M.2005.LandscapepatternsofexurbangrowthintheUSAfrom1980to2020.EcologyandSociety10(1):32.[online]URL:http://www.ecologyandsociety.org/vol10/iss1/art32/.
