PROCEDURESTO ENHANCETHESUCCESS OF A BLACKBEAR REINTRODUCTION PROGRAM FRANKT.VANMANEN,Departmentof Forestry,Wildlifeand Fisheries,The Universityof Tennessee, P.O.Box 1071, Knoxville,TN 37901, USA, email:
[email protected] R. PELTON,Departmentof Forestry,Wildlifeand Fisheries,The Universityof Tennessee, P.O.Box 1071, Knoxville,TN MICHAEL 37901, USA Abstract: Black bears (Ursus americanus)were extirpatedfrom the Big South Forkarea(BSFA) of KentuckyandTennesseearoundthe turnof this century. We developed a habitatsuitabilityindex (HSI) model to study the feasibility of a black bear reintroductioninto this area. We appliedthe HSI model to the BSFA and identifiedhigh- and low-qualityhabitatcomponents,their spatialdistribution,and their projectedchange over time. Potentialfor humaninfluences seems to be the weakest aspect of the BSFA as a successful reintroductionlocation. In conjunctionwith the habitat analysis, we developed a detailed release protocol, identifying factors that may furtherincrease the probabilityof reintroductionsuccess. Our approachcould apply to other bear species and geographicareas because of the flexible structureof HSI models and acceptanceof varioustypes of data. Int. Conf. Bear Res. and Manage. 9(2):67-77 Key words: black bear, habitatevaluation,Kentucky,reintroduction,release procedures,Tennessee, Ursus americanus.
Throughoutthe world, bear populations are affected by habitatloss andfragmentation,oftenresultingin population isolation, overexploitation,extermination,and increases in human-bearconflicts (Schoen 1990, Servheen 1990). Long-termsurvivalof isolated populationsmay be substantiallydecreasedthroughdemographicand environmentalstochasticities, naturalcatastrophes,or genetic effects (Shaffer 1987); reintroductionof bears into unoccupied range or augmentationof existing populations may be a valuabletool to counteractsuch negative effects (Griffithet al. 1989). Althoughthe blackbearis the most numerousbearspecies in the world,its currentdistributionin the southeastern U.S. is fragmented(Fig. 1) because of habitatloss and overexploitation,comprising less than 10% of the historic range (Maehr 1984). Recently, however, bear habitatrecoveryhas occurredin some regionsdue to land use changes and acquisitionof lands by state and federal agencies. Such habitatsoften cannotbe exploited by the species throughnaturalcolonization. Black bears were successfully translocated into the AtchafalayaandTensasRiverbasins of Louisianato augment populations(Taylor 1971). The reintroductionof black bears into the InteriorHighlandsof Arkansaswas likely one of the most successful reintroductionsof large carnivoresin the world (Smith et al. 1990). About 250 bearsfrom Minnesotaand Manitoba,Canadawere introduced in the 1960s to an areawherebearshad been extirpated;the currentpopulationas a resultof these effortsis estimatedat >2,500 bears(Smithet al. 1990). Otherthan the above, there has not been a concerted,systematicattempt at restoringa black bear population.
The BSFA of KentuckyandTennesseeis an areawhere land abandonmentand subsequentpurchaseby government agencies has createdpotentialhabitatfor bearreestablishment.Althoughblack bear sightingsoccasionally are reported,no residentblack bear populationhas been present in this area since the turn of the century. The closest residentpopulationof blackbearsis approximately 150 km away. To reestablisha black bear populationin this area,a reintroductionprogramwould have to be initiated. Reintroductionshave received growing interest as indicated by the formationof the InternationalUnion for Conservationof Nature and NaturalResources (IUCN) ReintroductionSpecialistGroupin the late 1980s (Stuart 1991). Reintroductionis defined as the intentionalrelease of individualsof a species into an areafrom which it has disappeared,with the goal of establishinga viable population(Stanley Price 1991). In reviewing the success of intentionalreleases of birds and mammalsfrom 1973 to 1986, Griffithet al. (1989) found that success of translocationsstrongly depended on the habitatquality of the release area. Moreover,translocationsof omnivorous species were found to be least successful (38%) comparedwith carnivores(48%) and herbivores(77%) (Griffithet al. 1989). The above findings, substantial projectedcosts for reintroduction,and the high profile of bearsjustify the use of habitatevaluationstudies before reintroductionattempts. However,success of a reintroductionwill not only depend on habitatquality but also on the release procedures,as was indicatedby the analyses of Griffith et al. (1989) and Maguire and Servheen (1992). The objectives of this study were to
68 Int. Conf Bear Res. and Manage. 9(2) 1997
Fig. 1. The Big South Forkarea (BSFA). Shaded areas represent current black bear distribution(Pelton 1982, Maehr1984). Study area sections are numbered 1-5.
(1) evaluate the feasibility of a black bear reintroduction in the BSFA by determiningrelativequalityof black bear habitat, and (2) develop a detailed release protocol. Fundingfor this study was providedby the Kentucky Departmentof Fish and Wildlife Resources; Tennessee Wildlife ResourcesAgency; the U.S. Departmentof AgricultureForest Service, Steams RangerDistrict, Daniel Boone NationalForest;and MclntireStennisProjectNo. 55 of the Departmentof Forestry,Wildlife and Fisher-
ies at the Universityof Tennessee. We thankR. Lozada, J. Bean, R. Thompson, and S. Bakaletz for assistance with field work. We sincerely appreciatethe help of the following people: B. Brum,J. Collier, R. Cornelius, T. Des Jean, J. Hersel, R. Hines, K. Lawrence, M. Melton, J. Thorsen, G. Weick, and P. Winistorfer. Special thanksgo to K. Johnson,F. Knauer,L. Marcum,C. McLaughlin, P. McLean, R. Pias, B. Stiver, and G. Wathenfor theircommentson the HSI model. We thank J. Clark for review of earlier drafts of the manuscript.
OFBLACKBEARREINTRODUCTION SUCCESS ENHANCING * van Manen and Pelton
STUDY AREA The Big SouthForkbasinof the CumberlandRiverwas in the southeasternpart of Kentucky and north-central Tennessee (Fig. 1). The study areaencompassedpartof Steams RangerDistrictof Daniel Boone NationalForest (DBNF), administeredby the U. S. Dep. of Agric. For. Serv., and the Big South ForkNationalRiver and RecreationArea (BSFNRRA), administeredby the Natl. Park Serv. (NPS) (Fig. 1). The study areawas about780 km2 and was divided into 5 sections to allow identificationof low- and high-quality habitat within the study area (Fig. 1). The climate of the BSFA was classified as humid mesothermal with little or no water deficiency (Thomthwaite1948). Annualprecipitationwas approximately 127 cm (Natl. Oceanographicand Atmos. Adm. 1978). Mean temperaturesrangedfrom -1 C to 10 C in Januaryand from 19 C to 32 C in July; the mean number of frost-freedays was 179 (Natl. Oceanographicand Atmos. Adm. 1978). The study area largely was within the mid and northern Cumberlandplateauregions of the AppalachianPlateau PhysiographicProvince (Thombury1965, Smalley 1986). The topographywas characterizedby long, narrow to moderatelybroadridgetopsand valleys with steep sideslopes (Smalley 1986). Averageelevationof the plateau areasrangedfrom 490 m in the southernportionof the study area to 395 m in the northernportion. The river gorge reached 180 m in depth and was the most characteristicfeature of the area. The diversity of abiotic conditions in the BSFA was reflectedby the relativelyhigh plant and faunaldiversity (Knowles et al. 1990). Bailey (1980) classified this area as theAppalachianoak andmixed-mesophyticforest sections in the easterndeciduousforest provinceof the lowland ecoregions. The forest vegetation in DBNF was managed for multiple use. NPS managementpolicies only permitted removal of timber for development or maintenanceof historic, public use, and administrative sites. Huntingseasons were in place for small and large game on the entire study area. The 1990 humanpopulation of all counties in the study area region was 87,946 or 13.2 people/km2(U.S. Bur. of the Census 1992).
METHODS HSI Model We developed an HSI model to estimatehow the relative quality of habitatvariablesand componentsmay affect the success of a black bearreintroduction.Ourblack
69
bear HSI model was based on a literaturereview, longterm researchdata, and experts'opinion. For a detailed documentationof the black bear HSI model and its verification level, see van Manen (1991). The HSI is definedas a numericalindex thatrepresents the capacity of a given habitatto supporta selected species and is based on the measurementsof habitatvariablesthatstandardizehabitatquality(U.S. Fish andWildl. Serv. 1981). The HSI is determinedthrougha combination of varioussuitabilityindex (SI) values for identified habitatvariables. An SI value representsa value of interest (the measuredhabitatcondition of the variable)relative to a standardof comparison (the optimum habitat condition of the variable) (U.S. Fish and Wildl. Serv. 1981). By definition, the HSI and SI values provide a 0.0-1.0 index of habitatsuitabilityfor a particularhabitat or habitatvariable,respectively(SchambergerandO'Neil 1986). A direct linear relationshipis assumed to exist between the HSI value and carryingcapacity(U. S. Fish andWildl. Serv. 1981). Harrisand Kangas(1988) arguedthatthe definitionof primaryhabitatbe extended for "gammaspecies" (species that depend on a regional landscapefor their existence) to includethe requirementthatan areahas sufficient size or configurationto supporta viable population. Because bears use habitat on a landscape scale (Schoen 1990), we chose landscape-scalevariablesassociatedwith primaryblack bearhabitatrequirements.Besides the assumptionof a linearrelationshipof HSI and carryingcapacity,3 more assumptionswere necessaryto implement the HSI model: (1) the SI curves of the variablesrepresent actualspecies-habitatrelationships,(2) where habitat variablesor componentsare dependent,relationships can be described through mathematicalequations, and (3) the entiremodel is used to evaluatean area. Black bear habitatuse in the southernAppalachianregion is associated with relatively large and undisturbed oak-hickory (Quercusspp.-Carya spp.) and mixed-mesophytic forests with abundanthard and soft mast production(Pelton 1982). The HSI model included8 habitat variablesrelated to 3 habitatcomponents: food, cover, and humanimpact (Fig. 2). We quantifiedhabitatvariables by use of the relationshipsbetween habitatmeasurementsand SIs describedin the HSI model (Fig. 3). Fourvariableswere used to calculateindices of summer food availability(V 1), fall food availability(V2), fall food productivity(V3), and fall food diversity(V4) (Fig. 3AD). The cover componentin the model was represented by 2 variables: protectivecover (V5) and potentialtree den sites (V6) (Fig. 3E-F). We determinedthat density readingsof >60%(accordingto Nudds[1977]) adequately
70 Int. Conf.Bear Res. and Manage. 9(2) 1997 Habitat Habitat Component
Habitat Subcomponent
Habitat Variables
Summer Food
VI-Coverpercentageof all softmastproducingspecies
Falld
V2-Relativecoverpercentageof all hardmastproducingspecies of all hardmastproducingspecies> 50 yearsold V3-Percentage hardmastproducingspeciesgroups V4-Numberof codominant
-NaturalProtective Cover HSI-
-Cultural
HumanImpact
V5-Coverpercentageof understory vegetationwith>60% vegetationdensity
TreeDen Sites
V6-Coverpercentageof hardwoodvegetation>150 yearsold
Roads Roads
V7-Open road length per unit area (km/km2)
Humanl Development
conflictzones of areain potentialhuman-bear V8-Percentage
Fig. 2. Schematic diagram of the structure of the black bear HabitatSuitability Index model for the southern Appalachians.
represented protective cover by sampling high density understory vegetation known to be used by bears in Great Smoky Mountains National Park, Tennessee. Open road density (V7) and human development (V8) represented the human impact variables (Fig. 3G-H); these 2 variables were included in the model to evaluate the potential for illegal hunting, human disturbances, and human-bear conflicts. The size of potential human-bear conflict zones was determined based on average female home ranges in the southern Appalachians. The SI values of the 8 variables and the component indices (CI) of the 3 habitatcomponents were weighted according to importance and compensatory relationships among the variables or components were used to calculate an overall HSI value: = (SIvI.(SIV2+SIV3+SIV4)2)1/3 CIFOOD 3
(1)
= SIV5 (2) = (SIV5+SIV6)else CICOVER If SIV5 ? SIv6, CICOVER 2 = (SIV7+ SIv8) CIHUMANIMPACT
HSI
+ CICOVER+ CIHUMANIMPACT) (2-CIFOOD =
njl
r)4 ^
(3)
(4)
Model Application All variablesrelatedto vegetation structureand composition were measuredin the study area in 1990. We collected vegetation data from 101 randomfield plots. We located each plot from U.S. Geol. Surv. 1:24,000 topographicmaps. At each location, 4 quarterswere establishedaccordingto the point-centeredquartermethod of Mueller-DomboisandEllenberg(1974). In each quarter of the sample sites, we estimated coverage of soft mast producingspecies with the releve methodof BraunBlanquet(1932). We determinedrelativecover and age of hard mast producing species by measuringdiameter at breastheight (DBH) of the sampletrees and collecting incrementbore samples (Avery 1975). We considered3 hardmastproducingspecies groupsto determinefall food diversity: white oak group, red oak group, and a group with hickory,Americanbeech (Fagus grandifolia), and black walnut (Juglans nigra). We estimatedvegetation density by determiningthe percentageof vegetationthat covered a vegetationprofile board (2-m height) at 15-m readingdistance(Nudds 1977, GriffithandYoutie1988). We estimatedavailabilityof potentialden trees (DBH > 84 cm; Johnson1978) andotherpotentialden sites within a 50-m radiusof each sample point. Open road density
* van Manen and Pelton ENHANCING SUCCESS OFBLACKBEARREINTRODUCTION
A.
.
.
.
.
.
.
.
.
.
.
.
.
.
B.
.
1.0
1.01 I
0.8
I
*
g 0.8I
0.6
.3 0.2
.
ca
*
i
*
*
0
. . . .
25
. . . . I . . . .
75
50
a
A
I
a
i
-
-
0.8
.
.
-
.
.
.
I
.
/
a
I
a
-
a
I
-
0.4
/
_
T .... 25
0.0 , 0
100
Coverpercentageof all soft mastproducing speciesin evaluationarea
~
I1
0.2
-
A u.un1 . . . .
1.Au
0.6
*
-
0.4
C.
71
50
75
100
Relativecoverpercentageof all hardmast producingspeciesin evaluationarea
D.
-
1-0' \
x
0.6/
\
0.4
fln C IIIaVIII v.v
I
0
E.
Orl
..
0.2
0.0
25
50
75
Percentageof all hardmastproducingspecies >50 yearsold in evaluationarea I a*
.
1
100
speciesgroupswithin studyareasections
.
I
2 ormore
Number of codominant hardmastproducing
1.0
^----------. 7T ~~~~~~~~0.8 0.26-0.4
_/. i
0.0
a
Ii
0
a.
I
25
I
I
I
I
50
[
I
I
i
75
100
Coverpercentageof understoryvegetationwith >60%vegetationdensityin evaluationarea -
.0 1.0
.
I
_111lilla
.I
.
.
I
.
I
a
a
K
0.
0
'1.0
I 00
I> 0.6
4. !
0.L 0.0r
lI
I 0
\
,
0.25
0.50
.
.
.
.
.
0.75
.
.
. -
....
1.00
Openroaddensityin km/km2 of evaluation area
50
75
100
-
--
0.88 I
\
10.4
C
I
....
0.6
I
0.
25
Coverpercentageof hardwoodvegetation>150 area H.-. yearsold in evaluation ' ' '
,,
0.2 I ..
0.0.
0
a I I I I I
25
50
I II1
75
100
Percentageof evaluationareain potential human-bearconflictzones
Fig. 3. Suitability indices (Sl) for 8 habitat variables used in the black bear HabitatSuitabilityIndex model. (A) Summer food availability.(B) Fallfood availability.(C) Fallfood productivity.(D) Fallfood diversity. (E) Protective cover availability.(F) Tree den site availability.(G) Open road density. (H) Human development.
was measured from U. S. Geol. Surv. 1:24,000 topographic maps. We also included unmappedroads that
we identifiedin the field. Potentialhuman-bearconflict zones were identifiedfrom field observationsand a visi-
72 Int. Conf.Bear Res. and Manage. 9(2) 1997 tor impact study (Turnerand Hammitt 1983). We used continuousinventoryof standcondition(CISC)data(U.S. Dep. of Agric. For.Serv. 1988) for DBNF to confirmand supplementthe field measurements. All measurementsand correspondingSI and CI values were calculatedfor the entire study area and each of the study area sections. We assessed differences in habitat suitabilityamong study areasections based on euclidean distances (Morrison 1990) among weighted suitability indices of the food, cover, and human impact components. The euclideandistances were used to constructa dendrogram.All statisticalanalyses were performedusing StatisticalAnalysis System software(SAS Inst., Inc. 1990).
Release Procedures We identifiedandevaluatedfactorsrelatedto 4 sources of variationthat could influence short-termreintroduction success: (1) survival,(2) reproduction,(3) post-release movements, and (4) human-bear conflicts. We applied a logistic regressionmodel based on Griffithet al. (1989) to find how habitatquality,numberof release animals,and length of programmay influence probability of reintroductionsuccess. We used this generalmodel to representlate-breeding,nativegame mammalsthatare released within the core of historicrange. This was not an attemptto extrapolatethe analysis of Griffithet al. (1989) to the species level or beyond the general conditions of their data. We identified the best release areas based on the HSI outcome. Otherfactors relatedto release procedureswere determinedbasedon previousfield researchand literature.
RESULTS HabitatEvaluation For each variable,measurementsare reportedfor the entire study area and ranges and standarddeviationsare based on studyareasections. Threemajorsoft mast producing genera(Vaccinium,Gaylussacia,andRubus)provided 7.3% coverage (range = 2.5-12.5%, n = 5, SD = 3.55), resulting in an SI for summerfood of 0.73 (Table 1). Mean cover percent of soft mast producers was significantlyhigher on the plateauareas (x = 9.60, n = 188, SD = 16.54) comparedwith the gorge (x = 5.19, n = 216, SD = 15.81) (Z = 9.59, P < 0.0001). Fall food producerswererepresentedby the generaQuercus,Carya, Fagus, and Juglans and comprised56.1% relativecover (range = 46.1-54.3%, n = 5, SD = 8.79). The absolute
frequency of hard mast producing species was 79.2%, resultingin an SI value for hardmast availabilityof 1.0 (Table 1). Mean relativecoverage of hardmast producing species was not differentfor the plateau (i = 3.56, n = 8, SD = 3.73) comparedwith the gorge (x = 2.51, n = 11, SD = 2.93; Z = 0.54, P = 0.59). Based on a linear regression equation and 95% confidence intervals (AGE = 24.35 + 1.21 x DBH; r2 = 0.40, n = 83, P = 0.0001) derived from the vegetation samples, we determined that at least 29.4% of the hard mast producing trees were older thanthe minimumproductiveage of 50 years. The associatedSI value for fall food productivity for the study area was 0.74 (Table 1). Codominanceof hardmast producingspecies groupsoccurredin 3 of the 5 study area sections, resulting in an SI value for fall food diversityof 0.80 (Table 1). Protectivecover was availableat 16.7%of the sample points (range= 10.7-24.0%, n = 5, SD = 5.49) resulting in an SI of 0.66 for this variable(Table 1). Vegetation density readings were significantly higher in the gorge (x = 45.77, n = 214, SD = 28.53) compared with the plateauareas(i = 36.14, n = 188, SD = 25.37; Z = -3.43, P = 0.0007). The field surveysdid not resultin the identificationof any potentialden trees. The CISC dataindicated the existence of one 19.8-ha old-growth Eastern hemlock (Tsuga canadensis)-hardwoodstandin DBNF. The lack of potentialtree dens resultedin an SI of 0.0 for this variable. Potentialrock dens were found at 26.7% of the sample locations. Density of open roads was 1.17 km/km2for the entire study area (range= 0.55-1.75, n = 5, SD = 0.44), which resultedin an SI of 0.0 (Table 1). The 912.5 km of open roadsin the areawere mostly graveland dirtroads;only 125.9 km of these roadswere improved. We identified7 potentialhuman-bearconflictzones;these zones included locally popularrecreationareas,developedvisitor facilities, developedcampgrounds,and concentrationsof residences. Five of these areaswere within the rivergorge. The potentialconflict zones comprised172 km2or 22.1% of the study area(range= 14.8-29.4%, n = 5, SD = 5.57), resultingin an SI of 0.29 (Table 1). The CI values for the food, cover, and human impact components were 0.81, 0.66, and 0.15, respectively, resulting in an overall HSI of 0.61 (Table 2). The euclidean distances among the study area sections based on CI values for the 3 habitat components showed that study area sections 1 and 4 were most similar, followed by sections 2 and 3, whereas section 5, the National Forest area, was the least similar to any of the other sections (Fig. 4).
SUCCESS ENHANCING OFBLACKBEARREINTRODUCTION * van Manen and Pelton
73
Table1. Suitabilityindices(SI)of blackbearhabitatvariablesforthe BigSouthForkstudyareain southeasternKentuckyand north central Tennessee, 1990.
indicesa Suitability
Studyarea section
VI
1 2 3 4 5 Overall
0.75 0.78 0.69 1.00 0.25 0.73c
V2
V3
V4
V5
V6
V7
V8
1.00 1.00 1.00 1.00 1.00 1.00c
b b b b b 0.74c
1.00 0.50 0.50 1.00 1.00 0.80d
0.80 0.43 0.59 0.96 0.50 0.66c
0.00 0.00 0.00 0.00 0.00 O0.00
0.00 0.00 0.38 0.00 0.00 0.00c
0.00 0.00 0.02 0.00 1.00 0.29c
a Key to variables: VI = Cover percentageof all soft mast producingspecies V2 = Relative cover percentageof all hardmast producingspecies V3 = Percentageof all hardmast producingspecies >50 years old V4 = Number of codominanthardmast producingspecies groups V5 = Cover percentageof understoryvegetationwith >60% vegetation density V6 = Cover percentageof hardwoodvegetation>150 years old V7 = Open road length per unit area (km/km2) V8 = Percentageof area in potentialhuman-bearconflict zones b Not calculateddue to small sample sizes. c Calculated separatelyfrom study area sections. d Average of all study area sections.
DISCUSSION HabitatEvaluation The resultsof the HSI evaluationindicatethe adequacy of food and, to a lesser extent,protectivecover. The limitingfactorswithinthe food componentwere summerfood availabilityand fall food productivity.If the SI for summerfood on DBNF increasedfrom 0.73 to 0.81, the averTable 2. Component indices (Cl) and overall black bear habitat suitability index (HSI)for the Big South Fork study area in southeastern Kentuckyand northcentral Tennessee, 1990. index Component Studyarea section
Fooda
Coverb
1 2 3 4 5 Overalle
0.86 0.76 0.72 0.94 0.59 0.81
0.80 0.43 0.59 0.96 0.50 0.66
Humanimpactc 0.00 0.00 0.20 0.00 0.50 0.15
a See equation 1. b See equation 2. c See equation 3.
d See equation4. e Calculated
separatelyfrom study area sections.
HSId 0.63 0.49 0.56 0.71 0.55 0.61
age for the BSFNRRA,the HSI would increaseby 1.6%. Projectionof futurehardmastproductivityindicatedthat, due to naturalmaturationof the forest, at least 48.8% of all hardmast producingspecies would reachthe optimal productionage withinapproximately10 years;this would result in a 3.3% increasein HSI to 0.63. Because black bearsuse securewinterdens otherthantreedens, we used treeden availabilityonly as a positivelycontributingvariable, andthe low SI values did not affect the overallHSI. In the long term,forestmaturationwill resultin increased availabilityof tree dens, which may increasethe HSI by 6.6% to 0.65. The potential influence of human activities was high relativeto the otherhabitatvariables. If the negativeeffect of human-relatedactivitieswere eliminated,the HSI outcome would increaseby 34.4% to 0.82. Therefore,to increasethe qualityof habitatin the shortterm, it seems most effective to managethe potentiallynegativeeffects of human activities ratherthan the naturalhabitatcomponents. Reintroductionsof large carnivoresoften will requirepublic support,as was suggestedby earlyred wolf (Canis rufus)reintroductionattempts(PhillipsandParker 1988) and grizzly bear (Ursus arctos) augmentations (MaguireandServheen 1992). Integrationof publiceducation andinformationprogramsinto bearreintroduction projectsprovides a way to identify and resolve potential conflicts before release. However, such programsmay
74 Int. Conf.Bear Res. and Manage. 9(2) 1997 Euclidean Distance 0.40
0.30
0.20
0.10
0
1
3 2 4 Section Area Study
5
Fig. 4. Dendrogramof Big South Forkstudy area sections in southeastern Kentuckyand north central Tennessee based on euclidean distances among weighted component indices of the habitatcomponents food, cover, and human impact of the black bear HSImodel, 1990.
need to be complementedwith a reductionof road density in certainareas and propergarbagemanagementat humanactivity sites. Study area sections 1 and 4 seems to provide the best black bear habitatbecause of high availabilityand productivityof food sources along with sufficientprotective cover. The availabilityof protectivecover was relatively low for sections 2 and 5. The potentialfor humanimpact seems to be the weakestaspectof sections 1-4. The quality of all 3 habitatcomponentsof section 5 was relatively low. The qualityof the summerfood componenton DBNF may be improvedthroughminor changes in forest management;soft mast availabilitymay be enhancedby prescribedburning(Harlowand VanLear 1989). Although the potentialfor human impact in sections 1 and 4 was relativelyhigh, these sections providedthe highest quality naturalhabitat. The centralposition and remoteness of these sections indicate the importanceof this portion of the studyareato establisha corepopulationfromwhich dispersalcould takeplace. The rivergorges arerelatively inaccessible to people and provide the best protective
cover for bears. It seems importantto maintainand enhance the remote characterof these areasto providepotential travel and dispersalcorridorsfor bears. As all habitatmodels, HSI models are simplifications of the systems they depict (Schamberger and O'Neil 1986). A good model preservesmost system dynamics (MaynardSmith 1974) in the simplestway possible. HSI models have been criticizedbecause of unreasonableassumptionsand oversimplificationof species-habitat relationships. However, the effectiveness of models depends on the intendeduse of the results (Laymonand Barrett1986). The simple structureof our HSI model was intendedto objectivelyand systematicallyassess the feasibility of a black bear reintroductioninto the BSFA; we were able to determinethe relativequality of habitat variablesand components,their spatialdistribution,and the effects of projectedhabitatchanges over time. Another criticism with regardto habitatmodeling in generalis thatmanymodels arenevervalidated(Stormer and Johnson 1986). We submitthat our HSI model has not been validated. The ultimatemeasureof habitatsuitability may be the ability of an area to sustain viable populations(Lanciaet al. 1986). Obviously,this is difficult because many factors not directly relatedto habitat (e.g., predation, competition, weather) may influence animalpopulations(Schambergerand O'Neil 1986). In case of extinct or remnantpopulations,little or no habitat use data can be acquired,thus limiting habitatevaluations to relatively general models based on the best informationavailablefrom otherpopulationswith similar habitatconditions. Due to the flexible structureof HSI models and acceptance of differenttypes of input data, our approachapplies to other bear species and different geographic regions. On a regionalscale, the HSI model may be useful to identify which areas need most protection,where populationscan be augmentedor reintroducedmost effectively, and how population linkages may be established. Although our model was not designed for use with GIS data, the rapid growth of this technology has greatly facilitated regional habitat modeling. Two importantfactorsin developingGIS-basedHSI models are: (1) compatibilityof resolutionof informationused in the model and in the GIS, and (2) the feasibility of generalizing habitat requirements so that GIS-based variables can adequately represent the life requisites of the species (Donovan et al. 1987). Bears are ideal candidates for habitat analysis with GIS due to their broad habitat requirements and the appropriatenessof landscape scales to bearecology (ClarkandvanManen 1993).
ENHANCING SUCCESSOFBLACKBEAR REINTRODUCTION * van Manen and Pelton
Release Procedures Based on the logistic regressionequationof Griffithet al. (1989), various combinationsof habitatquality,programlength, and the total numberof release animalsindicated that the effect of increased programlength on reintroductionsuccess was strongestwhen relativelyfew animalsare released, and vice versa;the effectiveness of releasing many animals over a longer time may be limited (Fig. 5). We suggest that a minimum of 40 bears released over 6-7 years is logistically feasible and provides a reasonableprobabilityof reintroductionsuccess. A viable demographicstructurein a reintroducedpopulation depends partlyon the sex and age structureof the release animals. Using existing dataandexperts'opinion to evaluatedifferentsex and age groups for grizzly bear augmentation,Maguire and Servheen (1992) found that 4-year-oldfemales providedthe best trade-offamongreproductivecontribution,probabilityof retention,andprobability of conflict. Translocatedbears may be motivated to returnto their originalhome ranges because of familiarity with that area or social relationships(Beeman and Pelton 1976). In a study of transplantednon-nuisance brown bears, Miller and Ballard (1982) found that nonreturningbears of both sexes were younger than returningbears. Thus,the motivationto returnmay increase of Probability Translocation
0.92
0.85
0.78
Numberof Releases
4
50 5
1
Program Length (Years)
Fig. 5. Relationship between number of animals released and program length on the probability of success of translocation. Based on logistic regression equation of Griffith et al. (1989). Remaining settings of the model were kept constant and were based on late breeding, native game mammals released in good habitat within the core of historic range.
75
with age. Mace and Haroldson (1984) suggested that subadultfemale grizzly bears may not have invested as muchas adultbearsin establishingandmaintaininghome ranges, presumablyreducing the need to returnto their areasof origin. These studies seem to indicatethatreintroductionsuccess may be enhancedby releasing young adult bears (i.e., approximately4-6 years) because they areless likely to exhibitextensivepost-releasemovements, therebyreducingthe probabilityof conflict. Because of greaterfemale reproductivecontributionto a bear population, a 2:1 sex-ratio of females to males may further increasethe probabilityof reintroductionsuccess. The areaof originof the release animalsshouldclosely resemblehabitatof the releasearea,especiallywhen adult animals are used (Servheen et al. 1987), and should be within the same subspecies range. Stiver and Pelton (1997) found that captive or nuisance black bears may not providea viable optionfor reintroductionor augmentation. Indeed,Griffithet al. (1989) foundthatwild-caught animalsfromhigh-densityandincreasingpopulationshad thehighestprobabilityof translocationsuccess;theyfound no associationbetweensuccess andimmediateor delayed release. However,the success of red wolf reintroductions in the southeasternU.S. has been partly contributedto the use of an approximately6-monthacclimationperiod where captive-bredwolves were graduallyfed a natural diet (Phillips and Parker1988). After unacclimatedreleases in Arkansas, some male bears were sighted and killed up to 435 km away from release sites, although manybears,mostly females, remainedclose (Smithet al. 1990). An on-site acclimationperiodmay enhancereintroductionsuccess by reducingpost-releasemovements. When wild bears are used, however, the risk of human habituationwill increase with time. A maximumacclimation period of 4 weeks may be appropriateduring which humanpresence should be minimized. To further enhance the probabilityof reintroductionsuccess, bears should be released during summer when food sources usually are abundantand predictable. Due to high annual variationof hard mast abundancein the southern Appalachiansand associated shifts in black bear home ranges (Garshelisand Pelton 1981), release in fall may not be a feasible option. Alternatively,the winter denning periodmay be used as an acclimationperiodby placing bearsin naturalor artificialdens;afterden emergence bears are sedentaryand will likely stay within the vicinity of the release site, particularlyfemales with cubs. Feasibility studies will not be able to predict exactly where and when unusual events may occur (e.g., crop damage, human-bearinteractions). Post-release moni-
76 Int. Conf Bear Res. and Manage. 9(2) 1997 toringwill be importantto assess or preventsuch events and to provide data to measure reintroductionsuccess (StanleyPrice 1991). Reintroducedpopulationsarerelatively easy to study due to initial small size and known sex and age structure;intensivemonitoringmay provide a unique opportunityto test the performanceof population estimatorsand indices. In summary,we propose the following release proceduresfor a reintroductionto establisha blackbearpopulation withinthe BSFA: (1) a minimumof 40 bears should be releasedof which approximately2/3 shouldbe 4- to 6year-oldfemales andthe remaindermales of similarage, (2) release animalsshouldbe obtainedfrom stable or increasing high-density bear populations in the southern Appalachianregion,(3) release animalsshouldnot have a known history of humanhabituation,(4) an acclimation periodof< 4 weeks maybe desirable,(5) theprimerelease sites shouldbe within study areasection 4 and the southernportionof section 1 wherehabitatqualitywas greatest, (6) releases shouldbe spreadout equally over a periodof approximately6-7 years, (7) captureof animals should occur in mid summerto allow release in late summer,(8) releasedbearsshouldbe intensivelymonitoredusing standardradio-telemetryproceduresandall monitoringshould be thoroughlydocumented,so that(9) successcanbe evaluatedbased on previouslyestablishedcriteria.
CONCLUSIONS The HSI model provided an objective and systematic approachto evaluateblackbearhabitatin the BSFA. The potentiallynegativeeffect of humanactivitiesseems to be the weakestaspectof the BSFAwith regardto reestablishment of a black bear population. The effects of human activitiesonbearsaredifficultto predictbecausetheypartly dependon managementactionsrelatedto a reintroduction attemptandpublicattitudestowardsbears. Evenif an area has suitablebearhabitat,manydifferentfactorswill influence the success of a reintroductionattempt.The development of detailed release proceduresshould enhance the probabilityof reintroductionsuccess. Reintroductionattemptsshould be planned as experiments(May 1991), andthey shouldbe usedto betterquantify the factorsthatinfluencereintroductionsuccess. With regardto the BSFA, an experimentalrelease of a small numberof bearswould help to evaluatethe effects of humanactivitieswhenno managementactionsaretaken;such experimentcan also be used to test the effectiveness of differentrelease procedures. The techniquesandconceptsused in studycould apply to other bear species and different geographic regions.
Althoughourpurposewas to studythe biological aspects of reestablishinga blackbearpopulation,we pointoutthat social, economic, andpoliticalaspectsalso will be important factors for successful reintroductionof large carnivores.
CITED LITERATURE AVERY, T.E. 1975. Naturalresourcesmeasurements.McGrawHill, New York,N.Y. Second ed. 339pp. R.G. 1980. Descriptionof the ecoregionsof the United BAILEY, States. U.S. Dep. Agric., Misc. Publ. No. 1391. 77pp. 1976. Homing of blackbears BEEMAN, L.E., ANDM.R. PELTON. in the Great Smoky MountainsNational Park. Int. Conf. Bear Res. and Manage. 3:87-95. J. 1932. Plantsociology: the studyof plant BRAUN-BLANQUET, communities. McGraw-Hill,New York,N.Y. 439pp. CLARK, J.D., ANDF.T. VANMANEN.1993. Geographic information
systems and black bear habitatanalyses. East. Workshop Black Bear Res. and Manage. 11:137-153. DONOVAN, M.L., D.L. RABE,AND E.O. OLSON,JR. 1987. Use of
geographicinformationsystemsto develophabitatsuitability models. Wildl. Soc. Bull. 15:574-579. D.L., ANDM.R. PELTON.1981. Movementsof black GARSHELIS, bearsin the GreatSmokyMountainsNationalPark. J. Wildl. Manage. 45:912-925. AND C. REED. 1989. GRFFITH,B, J.M. Scotr, J.W. CARPENTER,
Translocationas a species conservation tool: status and strategy. Science 245:477-480. , ANDB.A. YOUTIE.1988. Two devices for estimating foliage density and deer hiding cover. Wildl. Soc. Bull. 16:206-210. HARLOW, R.F.,ANDD.H. VANLEAR.1989. Effects of prescribed burningon mast productionin the Southeast. Pages 54-65 in C.E.McGhee,ed. Proc.South.AppalachianMastManage. Workshop. Univ. Tenn., Knoxville. L.D., ANDP. KANGAS.1988. Reconsiderationof the HARRIS, habitat concept. Trans. North Am. Wildl. Natur. Resour. Conf. 53:137-143. K.G. 1978. Den ecology of black bearsin the Great JOHNSON, Smoky MountainsNationalPark. M.S. Thesis, Univ.Tenn., Knoxville. 107pp. AND R.R. HANNAN. 1990. KNOWLES,B.J., J.N. CAMPBELL,
Cooperativeinventoryof endangered,threatened,sensitive, and rare species, Daniel Boone National Forest, Steams RangerDistrict. U.S. For. Serv., The NatureConservancy, KentuckyState NaturePreserveComm., KentuckyDep. of Fish and Wildl. Res. Winchester,K.Y. 169pp. ANDE.M. LUNK.1986. Temporal LANCIA, R.A, D.A. ADAMS, and spatialaspects of species-habitatmodels. Pages 65-69 in J. Vemer,M.L. Morrison,and C.J. Ralph, eds. Wildlife 2000: modelinghabitatrelationshipsof terrestrialvertebrates. Univ. Wisconsin Press, Madison. LAYMON,S.A., ANDR.H. BARRETr.1986. Developing and testing
habitat-capabilitymodels: pitfalls and recommendations.
* van Manen and Pelton 77 ENHANCING SUCCESS OFBLACKBEARREINTRODUCTION
Pages 87-91 in J. Verner,M.L. Morrison,and C.J. Ralph, eds. Wildlife 2000: modeling habitat relationships of terrestrialvertebrates. Univ. Wisconsin Press, Madison. 1984. Scope of work and MACE,R., ANDM. HAROLDSON. proposed study design: grizzly bear population augmentation.Rep. to U.S. Dep. Inter.,Fish andWildl.Serv., Grizzly Bear Recovery Coord. Missoula, Mont. 32pp. MAEHR,D.S. 1984. Distributionof blackbearsin easternNorth America. East. WorkshopBlack Bear Res. and Manage. 7:74. 1992. Integratingbiological L.A., ANDC. SERVHEEN. MAGUIRE, and sociological concerns in endangered species management: augmentationof grizzly bear populations. Cons. Biol. 6:426-434. MAY,R.M. 1991. The role of ecological theoryin planningreintroduction of endangered species. Symp. Zool. Soc. London 62:145-163. MAYNARD J. 1974. Models in ecology. CambridgeUniv. SMrITH, Press, Cambridge,U.K. 146pp. W.B.BALLARD.1982. Homingof transplanted MILLER, S.D., AND Alaskanbrownbears. J. Wildl. Manage. 46:869-876. D.F. 1990. Multivariate statistical methods. MORRISON, McGraw-HillPubl. Co., New York,N.Y. 495pp. 1974. MUELLER-DOMBOIS,D., AND H. ELLENBERG.
Aims and
methodsof vegetationecology. JohnWiley and Sons, Inc., New York,N.Y. 547pp. NATIONALOCEANOGRAPHICAND ATMOSPHERICADMINISTRATION.
1978. Climates of the States. Gale ResearchCo., Detroit, Mich. 1185pp. NUDDS,T.D. 1977. Quantifying the vegetative structureof wildlife cover. Wildl. Soc. Bull. 5:113-117. PELTON,M.R. 1982. Black bear. Pages 504-514 in J.A. ChapmanandG.A. Feldhamer,eds. Wild mammalsof North America; biology management, and economics. John Hopkins Univ. Press, Baltimore,Md. M.K., ANDW.T.PARKER.1988. Red wolf recovery: a PHILLIPS, progressreport. Cons. Biol. 2:139-141. SAS INSTITUTE,INC. 1990. SAS/STATuser's guide, Version6, Fourthed., Vol. 1 and2. SAS Inst., Inc., Cary,N.C. 1686pp. SCHAMBERGER, M.L., ANDL.J. O'NEIL. 1986. Concepts and constraintsof habitat-modeltesting. Pages 5-10 in J.Vemer, M.L. Morrison, and C.J. Ralph, eds. Wildlife 2000: modelinghabitatrelationshipsof terrestrialvertebrates.Univ. Wisconsin Press, Madison. J.W. 1990. Bear habitatmanagement: a review and SCHOEN, futureperspective. Int.Conf. BearRes. andManage.8:143154. C. 1990. The statusandconservationof the bearsof SERVHEEN, the world. Int. Conf. Bear Res. and Manage. Monogr.Ser. No. 2. 32pp. ANDA. CHRISTENSEN. , W. KASwoRM, 1987. Approaches to augmenting grizzly bear populations in the Cabinet
Mountainsof Montana. Int. Conf. Bear Res. and Manage. 7:363-368. SHAFFER,M.L. 1987. Minimum viable populations: coping with uncertainty.Pages 69-86 in M.E. Soule, ed. Viable populations for conservation. Cambridge Univ. Press, Cambridge,U.K. G.W. 1986. Classificationand evaluationof forest SMALLEY, sites in the NorthernCumberlandPlateau. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. SO-60. 74pp. ANDP.S. GIPSON.1990. History of SMITH, K.G., J.D. CLARK, blackbearsinArkansas:over-exploitation,nearelimination, and successful reintroduction.East. WorkshopBlack Bear Res. and Manage. 10:5-14. STANLEY PRICE, M.R. 1991. A review of mammal reintroductions,and the role of the Re-introductionSpecialist Groupof IUCN/SSC. Symp. Zool. Soc. London 62:9-25. STIVER,W.H., M.R. PELTON,AND C.D. Scott. 1997. Potential use of pen-reared black bears for augmentations or reintroductions.Int.Conf. BearRes. andManage.9(2):145150. STORMER, F.A., ANDD.H. JOHNSON.1986. Introduction: biometric approachesto modeling. Pages 159-160 in J. Verner,M.L. Morrison,and C.J. Ralph,eds. Wildlife 2000: modelinghabitatrelationshipsof terrestrialvertebrates.Univ. Wisconsin Press, Madison. S.N. 1991. Re-introductions:to what extent are they STUART, needed? Symp. Zool. Soc. Lond. 62:27-37. D.F. 1971. A radio-telemetrystudy of the black bear TAYLOR, (Euarctosamericanus)with notes on its historyand present status in Louisiana. M.S. Thesis, Louisiana State Univ., Baton Rouge. 87pp. W.D. 1965. Regionalgeomorphologyof the United THORNBURY, States. JohnWiley and Sons, New York,N.Y. 617pp. C.W. 1948. An approach toward a rational THORNTHWAITE, classificationof climate. Geogr. Rev. 38:55-94. D.A., ANDW.E.HAMMITT. 1983. Visitoruse andimpact TURNER, sites: the Big South Fork. Phase I: Observationalstudy. SoutheastReg. Natl. ParkServ.,Atlanta,Ga. 43pp. U.S. BUREAU OFTHE CENSUS.1992. 1990 census of population andhousing, southernstates. SummaryTapeFile 1A. U.S. Dep. Commerce. Washington,D.C. (Computerfiles). U.S.
DEPARTMENTOF AGRICULTUREFOREST SERVICE. 1988.
Continuousinventoryof standconditionsummary.U.S. For. Serv., Steams RangerDist., Whitley City, Ky. U.S. FISH AND WILDLIFE SERVICE. 1981. Standards for the developmentof habitatsuitabilityindex models. 103 ESM. U.S. Dep. Inter. Fish and Wildl. Serv., Div. Ecol. Serv., Washington,D.C. 162 pp. VANMANEN, F.T. 1991. A feasibility study for the potential reintroductionof blackbearsinto the Big SouthForkareaof Kentuckyand Tennessee. Tenn. Wildl. Res. Agency Tech. Rep. No. 91-3. 158pp.
