1170

U.S. Department of the InteriorU.S. Geological Survey

Arctic Refuge Coastal PlainTerrestrial Wildlife Research Summaries

U.S. Department of InteriorU.S. Geological Survey

Arctic Refuge Coastal PlainTerrestrial Wildlife Research Summaries

Edited by:D. C. Douglas, U.S. Geological Survey, Alaska Science Center, Anchorage, AlaskaP. E. Reynolds, U.S. Fish and Wildlife Service, Arclic National Wildlife Refuge, Fairbanks, AlaskaE. B. Rhode, Expression, Anchorage, Alaska

Biological Science ReportUSGS/BRD/BSR·2002·0001

U.S. Department of the InteriorGale A. Norton, Secretary

U.S. Geological SurveyCharles G. Groat, Director

U.S. Geological Survey, Reston, Virginia: 2002

Any use of trade, product, Of firm names in thispublication is for descriptive purposes only and doesnot imply endorsement by the U.S. Government.

Copks of this publication arc available from the National'Ii:chnical Information Service, 5285 Port Royal Road,Springfield, Virginia 22161 (I_HOO·553_6847 or 703-487­-1650). Copks also arc available to registatd users froillthe Defense Technical In[onJlation Center, Alln.: HelpDi:sk. 8725 Kingman Road, Suile 0944, Fon Belvoir,Virginia 22060-6218 (1-800-225-3842 or 703-767-9050).Digital copies of this publication C:l1l be obtained on theInternel at hllp:l/;ll;l.~ka.usgs.go\'.

('illllioll .'xamp!;':

Griffith. B., D. C. Douglas, N. E. Walsh, D, D. Young.T. R. McCabe, D. E. Rlls.~cll. R. G. White, R. I).

Cmwwn, and K. R. WhitteJ}. 2002. The Porcupinecaribou herd. Pages 8-37 in D. C. Douglas. P. E.Reynolds, and E. B. Rhode. editors. Arctic RefugecoaslaJ plain terrestrial wildlife research sUJlJmaries.U. S. Geological Survey. Biological RcsOIJrcesDivision, Biological Sciencc Report USGS/BRDIBSR-2002-0001.

ii

Preface

In 1980, when the U.S. Congress enacted the Alaska National Interest Llnds Conservation Act (ANILCA), it alsomandated a study of the coastal plain of the Arctic N~lliollal Wildlife Refuge. Section 1002 of ANILCA slated lha! acomprehensive invelltory of fish and wildlife resources would be conducted on 1.5 million acres of dIe Arctic Refugecoastal plain (1002 Area). Potential pctrolcUIll reserves in the 1002 Area were also (0 be evaluated from surface geologicalsludies and seismic exploration .~urvcys. Results of these studies and recommendations for [mllee management of theArctic Refuge coaS!3[ plain were to be prepared in a report to Congress.

In 1987, the Departlllent of Interior published the Arctic National Wildlife Rcfugt', Alaska, CDanal Plain RNDurceArsess!lwn! - Report and RCCDJIllllt'/ldlltirJ/l (0 tl/l' Congress o[ tl/l' United StaleS (lmi Final Elll'imnmclltal impactSWftmtlll. This report to Congress identified the potclIlial for oil and gas production (updated* most recently by IheU ,S, Geological Survey in 200 I), described the biological resources, and cvaluated the potential adverse effects to fishand wildlife resources. The 1987 report analyzed the Jlotelltial cnvironmental consequences of five managementaltemOltives for tlle coastal plain, nlllging frOIll wilderness designation to opening the entire area to IC3se for oil and gasdevelopmellt. The report's summary recollllllended opening the 1002 Area to an orderly oil and gas leasing program,but cautioned that adverse cffect.~ to sO/l1e wildlife Jlopulations were possible.

Congress did not act on this recoll\lllendation nor allY other alternativc for tlle 1002 Area, and scientists continuedstudies of kcy wildlife species and habitats on the coastal plain of thc Arctic Refuge and surrounding areas. This reportcontains updated summaries of those scientific investigations of caribou, muskoxen, predators (grizzly bears, wolves,golden eagles), polar bears, snow geese, and !heir wildlife habitats.

Contributiolls to tllis report were made by scientists affiliated with the U.S. Geological Sun'ey; U.S. Fish andWildlife Service; Alaska Departlllent of Fish and Game; Univcrsity of Alaska-Fairbanks; Canadian Wildlife Sen'ice;Yukon Department of Renewable Resources; and the Northwcst Terrilories Departmenl of Resources, Wildlife, andEconomic De,'c!opmenl.

Sections of the reJl0rt prcsclltillj,; ncw information on caribou and forage plants were peer-reviewed by thrceindcJlendent, non-affiliated scientists. The remaining sections sUlllmarize previously published peer-rcviewed scientificpapcrs and wcre reviewed by a single independcllC scicntist. The U.S. Geological SlIn'cy :md thc U.S. Fish and WildlifeServicc collaborated in the publication of this report.

:r U.S. Geological Survcy Fact Sheet FS-028-0lhup:1Igeology.cr.lls£s.g0vIpuh/fac t·.~hecIs/fs-002 ~ -0 II

iii

Pageiii

Contents

Preface.................................................. . ,.......... . .Authors' Addresses.................... . .Scclioll 1. Introduction. . .

Background.. . .Study ACI~a. 2References............................. 2

Section 2. Land Cover.......... "Vegct,ltiorl 1\1 appillg of the Arctic Refuge Coa.'>tal PI ~lin.. .j

References......................................... 6Seelion 3. The Porcupine Caribou Herd........................ 8

Data, Methods and Assumptions.................... SNUlritiorta[ Irnponance of the Calving Ground IIHabitat Trends During the SlIIdy Peliod... IIHerd Dynamics and Delllogfilphy...................... . 13Season",1 Distribution and I\lovell1ents.................. . 15Foraging on the Calving Ground.......... . 20Habitat Selection..................................... . 21Effects of Insect Harassment on Habitat Usc...... . 24Calf Performance in Relation to Habitat Usc........ . 25Factors Associated with Calf Sun'ival on the Calving Ground............... . 25Potl;ntial Effects of Dcvelopment on Junc Calf SurvivaL......... . 28Conclusions......................................... . 32Rcferences............................................. . 34

Section 4. The C('lltral Ardic Caribou Herd . 38Status of the Central Arctic Herd............ . 39Development-related Ch'lllges in Distribution...... . 39Body Condition and Reproductive PerfOnll;\llce.. .. 41Ovcrview.. . 43Referencc"'............................................. ... 44

Section 5. Forage Qualltit~' and Quality . . 46Forage Comparisons Within and Outside the 1002 Area.. 47References ,............... 50

Sectioll 6. Pr('c111tors....... . 51Prl:dator Distribulions................. 51Factors Associated with Prl:dator Distributions 51Rates of Prl:dation ".. 53References........... 53

Section 7.l\1uskoxen................ . 54Dynamics and Rangc Expansion of a Re-cstablishl:d /l.llISkox POPU1:ltion.. .. 54SC:lsonal Stratcgies of Muskoxl:n: Distribution, l-Iahit:\IS, and Activit}' Pattl:nls................ . 56Winter Habitat USl: of Muskoxen: Sp:llial Scales of Resource Selection....... 61Sulllmary......... . 62Rdercllccs......... 63

Sectiun 8. Polar Bears......... 651>1ovcnlCnts and Population Dyn'llllics of Polar Ilears ,... 65Reproductivl: Significancc of 1\lalcrnit}' Denning on Land 67Refcl"l:nccs.. 69

Section 9. Snow Gccs~. 71Sizl: and Distribution of Sllo\V Geese Populations 71Snow Goose l-Iabitn!. Food. and Energy Rl'q\lir(~rnents... 71Effect of l\li!ig,llion of I·hllnan Activitk, on Sno\\' Geesl: 73Refl:rl:llCl:s... 74

Acknowlcdgclll('lltS , ,........ . 75

FiguresNllmba Page1.1. Geographic 1l13Jl of the Arctic Nation;"!] Wildlife Rduge. Alaska, USA, fwd surrounding areas 1

1.2.

2.1.

3,1.

3.2.

,

i3.3.

,

'i3.4.

3.5.

3.6.

3.7.

3.8.

3.9.

3.10.

3.1 I.

3.12.

Geographic map of the 1002 Area of the coastal plain of the Arctic N:lIional Wildlifc Refuge, Alaska 3

L3nd-cover lIlap of the 1002 Area with corresponding vegdation class names, descriptions, and cklsscodc.~, Arctic National Wildlife Refuge, Alaska.......................................................... 5

Land·cover classes on the Arctic Refuge coastal plain and eastward as generalized for cariboll studies ...... 9

For the Porcupine caribou herd: anllual range, calving sites, and aggregate extent of calving, 1983-2001. For the Central Arctic caribou herd: calving sites and extent of calving, 1980-1995 .... ... 10

ivkan temperatures for 2 stations within the Porcupine caribou herd's aggregate extent of calvingand 1 stalion within its winter range for a) June and lJ) winter, 1950-1995.. 12

I\fcdian Normalized Difference Vegetation Index (NDVI) Oil 21 June within the aggregate extent ofcalving for thc Porcupine caribou herd, 1983·2001.............................. 12

Standardized values of the Arctic Oscillation (AO) for winter and standardized JlOJlUI;llion size of thePorcupine caribou herd, 1958-2001........................... 12

I\Iedian NDVI at calving within the aggregate extent oft:alving of the Porcupine caribou herd for thccurrent year, and winter Arctic Oscillation index for the previolls calendar year, 1985-2001..................... 13

•Frequency of days with daytime temperatures above freezing in a) spring and b) fall, on transitionalranges of the Porcupine caribou herd during the increase and decrease phases of the herd 13

Size of the Porcupine caribou herd, 1972-200 J , l'stimated frolll aerial photo-cl'l1.'ill.ses by the AlaskaDepal1rnent of Fish and Game.................. 13

Rel:ltive post-calving herd size.s of the" Alaska barren-ground caribou herds, 1976-200 I............ 14

Reproducti ve e.~tiJl1ates for the Porcupine caribou herd, 1983-200 I: (l) parturition ratc of adullfcmales, /I) calf survival from bit1h through the last week of June, ;lI1d c) net calf production 14

DistribUlion of sate II itc-collared Porcupine caribou herd felllales during 7 time periods, 1985-1995 J6

Minimulil median daily movcment rate of panuriellt satdlite-collared females of the Porcupinecaribou hcrd, 1985-1995......................... . , 17

3.13. Calving distributions of the Porcupine caribou herd, 1983-2001 . IS

3.1·1. Percellt ofrallio-collarl'd Porcupinc (ariboll herd females thaI (;llved in lh.: 1002 Ar~';t. 19RJ-2001 II)

3.15. )lerccllI of radio·collared Porcupine Ctrihou herd fem;tks lhat calved within tile 1002 Arl~a in rdatjrmto the median NDVI at calving within thl' aggregate extcnt of calving, 1985-2001.... 19

3.16. Porcupine carib\)lJ herd tI) diet cOll1pn.~iti()1l and II) median phenology of m;tjor foragL' itL'lllS, 19~n..... 20

3.17. Porcupine caribou herd ill diet composition and h) median phenology oflll<ljor forage items, 1994 20

3.18. Annual conditions of srWWcover and veget:llion phenology derived from saId lite im;lgcry duri ng thecalving period, 1985-2001, for the Porcupitw caribou 11l'rd 22

"

Figures (continued)Number Page3J 9, Average percent of area in low (,::; median) or high (> median) cla~se$ of a) daily rate of increase

in the NDVI, b) NDVI at calving, and c) NDVr on 21 June for the aggregate extent of calving,anllu;l1 calving grounds. and concentrated calving areas of the Porcupine caribou herd, 1985·200 I 23

3.20. Average percent of area in 4 exclusive Slloweovcr classes for the aggregate extent of calving,anllual calving grounds, and cOllCClIlratcd calving arca.~ of lhe Porcupine caribou herd, 1985-2001 ". 23

3.21. A\'l:ragc percell! of arca in 6 vegetation types for the aggregate ex lent of calving, annual calvinggrounds, and COllccrtlralcd c;llvillg arC;h of the Porcupine caribou htrd, 1985-200 I ....... ..... 24

3.22. Estimated total int;lke of diemry nitrogcll from the ctlving ground for·1 Nonh American e;lribou herds.". 24

3.23. Daily weight gain of caribou calves of thc Porcupine herd. 1992·199·;, during 2 timc periods 25

3.24. Availability of 6 vegetation types in the aggregate extent of calving f(lr the Porcupine cariboll herd andusc by radio-collared calves during 2 time periods .for 11) 1992. h) 1993, and c) 1994 26

3.25. l....kdian NDVI on 21 June within tlle Jnllual calving grounds of the Porcupine caribou herd and weightsofpartUrlellt female ehribOIl when captured withillthe annual calving ground 01\ 21 June. 1992-1994 ....... 26

3.26. Calf survival through June for the Porcupine caribou herd. 1985-2001, in relation t<) median NDVIon 21 June within the aggregate extent of calving.... . 27

3.27. Predicted calf survival through June for the Porcupine caribou had, 1985-2001, ill relation 10 Illcdi;lnNDVI 011 21 June withilllhc annual calving ground ;lnd to the proportion of calves born 011 the coastalplain physiographic zone whae prcdalOr dcnsity was lower than in the foothill-mountain zone 28

3.28. Estim;lted change in calf survival during June for the Porcupine caribou had, 1985-2001, as a functionof the distance of displacemcllt of the aBllllal calving ground and associated conccntrated calving arcaand calving sitc.~........ "............... . ".. 31

3.29. Aggregate c'.xtent of annll;11 c,lI ving and aggrcgate exten! of concentrated calving for the Porcupinecaribou herd, 1983·2001......................... . " " 33

4.1. Oil field infrastmcture in the Prudhoe Bay and Kilparuk !)('trolcUlll dcvelopment areas, Alaska, 1994.... 38

4.2. Photocellsus estimates of the Central Arctic <:aribou herd, 1978-2000, and net calf production based 011

observations of radio-collared adult females from 10 June through 15 August.. 39

4.3. Fractional changes in mean density of caribou from the Ccnlr;ll Arctic herd hetwcen prc-constl1lction(1978-1981) and post-constnJclion (1982·1987) periods for I-klll-diswnce intervab from thc Milne Pointroad system in the Kupamk petroleUlll dcvelopment area, Alaska.... . ,...... . ,...... 39

4.'1. Ch;lllges in mean relative distribution of c;lribou from the Central Arctic herd in the KUp;lnJk petroleumdevelopmcllt ,Irca, Alaska, during calving: 1979-1981, 1982-1986, and 19R7·1990 40

4.5. Decline in percent ;lbuIHlancc~ ofcariboll frolllthe ('L'lllr:il Arctic herd \\"('"t of the I\lilne POlut Roadand changes in IOtal numbers ~)f c,lriboll observed north of the Spine Road, 1979-R7. -10

4.6. Helationship between llle:lJ1 density of caribou from the Central Arctic had and road density withinpreferred mgged terrain, Kllp:lmk PL~trok1Jlll del'c!l)]llllL'Jlt ;lre:l, Absk:l, 1987-llJ92... . ,11

4.7. Shifts ill CO/lf,:;cnlmled calving areas. Central Arctic carihou Ilerd, Alaska, 1980-1995...

VII

41

Figures (continued)Nllmber Page4.8. Logistic regressions of parturition ralt::, incidence of carly calving, and perinatal calf 5urviv::11 on ;\ulumn

and SlImJlll'r body weights of female caribou, Central Arctic herd, 1987-1991............ 42

4.9. Changes in median NDV] on 21 1une for concentrated calving areas of the Celllf'l] Arctic caribou herdin the reference zOllc(rdativdy undisturbed) and treatment zone (active oil fields), 1985-1995 42

·~.IO. l\kan body weights of !;lcl:lling and llonbcl:lling female caribou from the CCllIr:1l Arctic herd in summerand autumn ,......... 43

4.11. Disiriblllion~ of Ilbs~rv~d <llltunlll body \\'~ight~ for lactating and Ilonlactating female caribou frol1lth~

CClIlral Arctic herd ."." "........... . ".......... "........... 43

5.1. 1\1<l]J of caribou fmagl.: study are<l, Porcupine caribou had, on th~ coastal plain of the Arctic Refuge. 46

6.1. Distribution of a) goJd~n eagle n~st structures, h) wolf dellS, and c) grizzly bears ncar the calvinggrounds of the Porcupine caribou herd......... . <" •••••• 52

7.1. Number of lllu..,koxell observed inlhel002 Area of the Arctic Refuge, 19S2-2001 54

7.2. Changes in r<lles of ~ucce~~ful production of Illuskox cal\'l~~ in thc1002 Area, Arctic Refuge, 1983·2001...55

7.3. Number of muskoxen killed or scavenged by grizzly bears from April 19H2through June 20tH innortheastern Alaska 55

7A. Range expansion of llluskoxcn in mixcd·.,ex groups in and near the Arctic Refuge, 1969-1993 57

7.5. Season<ll changes in ratl'~ of movemcnt and activit), CDums of satellit~-collaredfemale 11l11.~koXCll

in and ncar thc Arctic Refugc, 1986-1992 . 58

7.6. Locatiol1s of mixed-sex groups of muskoxen secn during wintcr and summer SUf\'eys in the ArcticRefuge, 1982-1999..... . 59

7.7.

8.1.

SIlOW depth in mll.~knx feeding zones, adjacent zones, and nonadjacent zoncs in late winter. 1989 and1990 OIl the cO;lslal plain of the Arctic Refuge.... . .

Numbers and relocation positions of satellite radio-collared Jlolar bears c<lplured in each of 6longitudinal z()rw.~ within the Bcaufort Sea .

61

66

8.2. Number of pol:!!" hear dens located hy radio-telemetry in c;\ch of 3 substrates. 1981-1990 ".... 68

8.3.

SA.

9.1.

/l.latL'mal den !<ll'ations for 5 ]lolar bl~ars followed to dens for llIore th;lll one )"e<lr. .

Distribution of lll;ltern;rl dens of radio-collared polar hears along the northern coast of Alaska andCanada, 19:-11-2001. . "........................... . .

NII!l1ber.~ or lesser SIll)\\, gcese observed on the cO;lsl<r1 plain of the Arctic Refuge, 1982:.1993 .

68

69

71

J

9.2. Frequency of lise of 25-kll1~ cclls by lesser snow goose /locks on the coastal plain or tile Arctic Rd"ugc,Il)82-llJ93. . ".......................... 72

lJ.3. Rat!:s of lipill dcpllsition by lesser snow gee.~e during fall staging OJ} the Arctic Refuge 73

viii

Tables

NI/mba Pagl:2.1. Contingency table used to assess the accuracy of the land-cover map of the coastal plain of the Arctic

National Wildlife Refuge. Alaska , " 6

2.2. Percent of each hmd-covCf class in the land-covCf lIlap of the coastal plain (If the Arctic Rdugc, andthe perce lit partitioned among various terrain Iypes...................... 7

3.1. Number of cllving siles, !lumber of calving Sill'S in the concentrated calving arca (CCA), an.~a of CCA,,\rca of annual calving ground (ACG). ralio of sizes of CCA 10 ACG, and population size of thePorcupine caribou herd, 1983-2001... 19

4.1. Parturition status of 43 raJio·<collared female <caribou. Central Arctic herd. wcSt and east of theSagavanirktok River, 1988-1994....... . 43

5.l. I'vkdian phenological stages of major forage spccics in the Por<cllpine <cariboll herd's traditionalcalving area and potcntial displaccmcnt art~a...... . 47

5.2. I'vledian biomass ,Illd pr:rCCl1t covcr of 4 major caribou foragc specics in the 5 most COllllllonvcgetationtypes 011 thc coastal plain during thc Porcupinc <caribou hcrd's calving pcriod, 1990 48

5.3. i\ledian dcnsity, bioll1ass, and pcrccnt covcr of major forage sJlccics in the Porcupinc cariboll herd'straditional calving arca and potential displaccmcnt arca, 1990·1991.... . 48

5.4. Medianllutricllt and fiber concentrations of 2 major forage species ill different phonologic:ll stages inthe Porcupine caribou herd's traditional caribou calving area and potential displacemellt area . 49

5.5. Distribution of vcgetation types in the caribou forage sludy area based on ,Hl illdependem sampleof 756 systematically-located vcgetatjon plots 49

5.6. I'vfedian lllltricJH and fibcr cOIl<ccntratiolls of tussock co!tongrass intlorcsccllccs in di [fercnt pherlOlogicalstagcs <comparcd betwecn tussock tundra and other vcgetation typcs... 50

7.1. Numbcr of lllll.~k()xcn scen in diffcrent rcgions in northeastern Alaska, USA, aud northwestern Cllladain19S2-2000during prc·cal"ing sun'cys 58

lX

Robert (Skip) Ambro..•USFWS Endangered Species

EcoIO\Ik:aI Services Field otrlCe-Fa!ttlanks101-12" A~IlIlUtl, Box 19Faltbanb, Alash 99701

Steven C. AmatrupUSGS Alaska ScleI'lC1l Cenler

1011 E. TudofRoadAnchoolge, Alaska 99503

AI.n W. BrackneyUSFW$ Arctic NalioMI Widlife Refuge

101 12" Avenoo. Room 236FIl!lba~S, Alaska 99701

Rlymond D. C.meron'Alaska Oepartmem of Fish and Game

1300 Colleg~ RoadFairbanh. Alaska (/9701

David C. DougluUSGS Alaska SClonce Cenler

P.O. Box 240009Oooglas, Alaska 99824

H.ney A. F,llxUSFWS Arctic Naliooal WI(lllfe Reruge

10112" Avenllll Room 238Fa~nks, Alaska 99701

G,r.ld W. G.m,r'USGS Alaska Science Cenler

1011 E. Tud(l(Roadmoorage, Alaska 99503

Bl1Id GrlfflthUSGS Alaska Cooperative

Fish and \'{idlife Research Unit209 Irving I Bukllng, P.O. BOJl mow

University or A1aska.fairbanbFairbanks. Alaska 99775

JerryW. HuppUSGS Alaska Science Cemer

1011 E. Tudor"RoadAnchorilge. Alaska 99503

• Current addresses:

Authors' Addresses

Jan,t C. Jorg.nlonUSFWS Altlk: National WMlife Reto;e

101·12" Avenue, Room 236Fairbanks, Alaska 99701

P,tu C. Jorl,'USGS Alaska Science Center

1011 E. Tudor RoadAnchorage, Alaska 99503

Oavld R. Kr.ln'USGS A1as~a Cooperative Fish and

Widl~e Research Unit2091rvlng Bulding

UrWelOiIy 0' AJaska-FallbanbFairbanks, Alaska 99775

Thom.l. R. NeC.b,'USGS Alaska SCience Cenler

1011 E. TudOfRoadAnchorag~, Alaska (/9503

D.n J. RttdAlaska Oepartment of Fish and Game

1300 CoIIeg~ RoadFairbanks, Alaska (/9701

H.rry V. R,ynord.A1aslr.a Oepartment of Fish and Game

1300 College RoadFairbanks, Alaska 99701

P.trlcl. E. R.ynold.USFWS Arctic National IIndlire Refuge

101·12" Avenue. Room 236Fairbanks, Alaska !19701

Donna G. Robertlon'USGS Alaska SCience Center

1011 E. Tudor RoadAnchorage, Alaska (/9503

Dould E. RuneltNoMem ConsllfVelion

Canadian W~dlife Service91782 Alaska Hi;lhway

WMehorse, Yukon Tenilorles Y1AS87Canada

W.lttrT.SmlthAJasi:a Department or Fl:!t11ltld GatM

1300 CoII&g1l RoadFaJrtnws. Alask.a 99701

'hrlo; S. UdivitzUSGS Alaska SCiellU Cenler

1011 E. TudorRMdAnchorage, Alaska 99503

NorMn E. W.tlh'USGS Alaska Science Center

1011 E. TudOf RoadAnchorage. Alaska 99503

Grtg J. W,II,r'USFWS Arclic Nallonal \'{Qdlil'o R~fuge

10112'" Avenllll, Room 236Fairbanks. Alaska (/9701

Robtl1 G. WhittlflSt~ule 01 Arctic Biology

University or A1aska-Fairbanks902 Koyukuk Ome

Irving I Buading, Room 311Fa~, Alaslr.a 99701

K,nn,th R. Whltt.n"Alaska Department of Fl:sh end Game

\300 CoIIi!ge RoadFairbanks, Alaska 99701

K,nn,th J. Wll.on'USGS Alaska Cooperative

Fish arid Wldlife Research Un~

University of Alaska-FalrbanbP.O. Bo.1: 757020

Fairbanks, Alaska !19775

Don.ld D. Young"USGS Alaska SCience Center

1011 E. TudOfRoadAnchorage, Alaska 99503

Robert Ambrose, National Park service, WASO, NRPC, Foil Collins, CORaymond D. Cameron. lnstnule or Arctic Biology. University or A1aska.fairtlanks, 902 Koyukuk Drive. Irving I Buiding. Room 311, Falfbanks. AK 99701Gerald W. Gamer, doceasedPeterC. JOM, U,S, GooIogicai Survey. Upper Mid'Ir\1st Environmental Science Center, 2630 Fama Reed Road, La Crosse. W! 54603David R. Klein, Inst~ule 01 Arctic Biology, Unlversit)' of A1aska·Fairbanks, 902 Koyukuk Drive, INio:!lBuilding, Room 311, Fairbanks. AI( 99701Thomas R. McCabe. U.S. Fish and Wildlife Service, EcoIog1cai Sef'lices, Chesapeake Bay Field Office. 177 AdmIral Cochran Or.. Annapolis. MO 21401Donna G. Rooollsoo, Halllirlg.lawson and Assoclales, 601 E. 57" Place, Anchorage. AI( 99518Noreen E. Walsh. U.S. Fish and WUdlifll Service, Ecological Services, Atlanta Field Office, 1875 Century Boulevalll. Room 200. Atlanla, GA 30.345Greg J. Weiler, U.S. Fish and Wildlife Service, Potomac National Wildlife ReltJge. 14344 JeNel!>on Davis H9hway, Woodhridge, VA 22181Kenneth R. Whitten. P.O. Box 81743. Faifbanks. AI( 99708Kennelh J. Wilsoo, A1yeska Pipeline Service Company, P.O. Box ro469. Faifban~s, AK !19701Donald D, Young. Alaska Department or Fish and Game, 1300 College Road, Fairbanks, AI( 99701

x

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL \VILDLIFE RESEARCH SUMMARIES

Section 1: Introduction

Background

The Arctic National Wildlife Refuge in northeasternAlaska is one of 16 refuges in Alaska and 539 refugesnationwide within the National Wildlife Refuge Systemildministcred by the U.S. Fish and Wildlife Service. Firstestablished as the Arctic National Wildlife Range in 1%0by Public Land Order 2214, it initially had a three-foldpurpose 10 preserve unique wildlife. wilderness, andrecreation values on 8.9 Illillion acres.

In 1980, the Arctic National Wildlife Range wasexpanded to the southwest and renamed the ArcticNational Wildlife Refuge (also called the Arctic Refuge inthis report) when the U.s. Congress passed the AlaskaNational Interest L~H1ds Conser\'ation Act (ANILCA),Public Law 96-487 (94 Stal. 2371). This legislation alsodesignated almost ,Ill of the original Arctic NmionalWildlife Range as wilderness, and it directed theSecret:lry of the Interior 10 conduct studies evaluatingboth the biological resources and the potential petroleum

reserves of 1.5 million acres (titled the 1002 Are:l) on thecoastal plain of the Arctic Refuge.

In April 1982, the Arctic Refuge staff completed areport summarizing the then current st:l1e of knowledgeon the fish, wildlife, and their habitats present 011 thecoastal plain of the Arctic Refuge (U.S. Fish and WildlifeSen'ice 1982). From 1982 to 1985, field investigations ofbiological resources of the 1002 Area were carried out bya number of investigators, and annual reports summarizedthe results (Garner and Reynolds 1983. 1984. 1985, 1986.1987). These reports and other resources were used toprepare a DepartlllelH of the Interior report to Congress:Arc/ic Na/ional Wildlife Rtfllge, Alaska, Coastal PlainRt'SOIlrce ASSt'SSIllt'1l1 - RI!J1orf (///(1 RecOIlllllclUiarioll toIhl' COIlgr,'ss of the Unired Slall'S and FinlllEIII'irolllllt'lIIal I/Ilpact Srart'/II<'1l/ {Clough el al. 1987).

Biological investigations continued from 1988 through1994 in and near the 1002 Area coordinated by researchscielllists from the U.s. Fish :md Wildlife SCr\'ice who arcnow part of the U.S. Geologic:ll Survey (~kCahe et a!.1992). Coll:lboralOrs included speci:llists from the ArcticNational Wildlife Refuge: Alaska Department ofFish andGame: Univc~ity of Alaska-Fairbanks: Canadi:ln Wildlife

- Amkl'<"\.\'I{boo<>d¥y

1002 bound.ry

-, In""""'>tionol~

Arctic National Wildlife Refuge {

and Surrounding Areas 'fi'"mil..

o 20 4ll 60 !lO~

031(,.4961:18

lilomo<=

Figure 1.1. Geographic map of lhe Arctic National Wildlife Refuge. Alaska, USA, and surrounding areas.

2 BIOLOGICAL SCIENCE REPORT USGSfDRD 2002·0001

Service; Yukon Department of Renewable Resources; andthe Northwest Territories Department of Resourccs,Wildlife, ;jlld Economic Dcvelopmenl. Additionalinfonnation continued to be collected from 1995 until thepresent (200 I) as part of monitoring caribou (Rallgi[utarandus), polar bear (Ursus maritimlls), and muskox(Ovibos moschatus) populations and their habitats.

This current report is a summary of these tecentinvestigations, building upon the information of paststudies. It includes updated information about populationdynamics, distribution, energetics, and habitat use of thekey wildlife species as well as discllssions about Jlotentialeffects and mitigation of petroleum development onwildlife and habitats in the 1002 Area.

Study Area

TIle studies summarized in this report focused on the1002 Area of the Arctic National Wildlife Refuge bllt alsoextended into adjacent regions of the Arctic Refuge,eastward into C<lnada, and as far west as the Prudhoe Bayand Kuparuk petroleum development areas in northcentral Alaska (Fig. 1.1).

'nte Arctic National Wildlife Refuge is the largest andlllOSt northern national wildlife refuge in the UnitedSlates, encompassing 19.6 million acres (30,000 squaremiles). A vnriety of arctic and subarctic habit<lts exist inthe Arctic .Refuge. including ncar shore marine habitatsalong tlte coast, arctic tundra on the cO:lstal plain, alpinehabitats in the foothills and mountains of the BrooksRange, and taiga and boreal forests south of themountains (Fig. l.l).

The cO<lstal plain of the Arctic Refuge containscalving grounds of the intemational Porcupine caribouherd, year-round habitats for muskoxen, fall staging areasfor lesser snow geese (Chell cacrulescellS c(/t'rllh'sccIlS).denning habitat for pregnalll polar bears, and SUllllllerncsting areas for numerous species of migratory birds,

TIle 1002 Area is a region on the coastal plain in thellorthem part of thc Arctic Refuge (Fig, 1.2). It liesbetween the mountains of the Brooks Range (69(> 35' N)and the Beaufort Sea (70° 10' N) and is bounded on thecast by the Aichilik River (l42°]{)' W) and on tIle west bythc Canning River (1460 IS' W),

Numerous northward-flowing rivers <lnd streamsbisect the 1002 Area. Only a few large lakes arc presellt:llld lll(}.~t freeze to the bOllom by latc winter, 'l1le climateis characterized by extremely low winter temperatures,persistent winds, and short cool SUlllmers, Temperatures at .Kaktovik on the coast of the Beaufort Sea (Fig. 1.2).averaged -25"C (-IYF) in FehnJary and +(j"C (+4J"F) inJune during 19H6-l995.

Precipitation occurs frequently as drizzle in SUllllllerand light snow in \\"iJ1l(~r. The ground surface is frozenfrom Scptember until Junc. A permanently frozcn

substrate called permafrost lies below the surface of thesoil. Winter conditions with below freezing temperaturesand snow exist for 8 to 9 months each year, Easterlywillds predominate most of the year, although stormsusually arrive on westerly winds, At K<lktovik, the sun iscontinuously above the horiwn from thy IS to July 27and below the horizon from November 24 to January 17.

The mountains of the Brooks R<lnge converge with lheBeaufort Sea in this northeastern corncr of Aillska, Theresult is a uniquc juxtaposition of landscape features inthe Arctic Refuge compared with surrounding arcas (Fig.1.1). The steeper elevation gradient between mountainsand ocean on the coastal plain condenses a diversity ofhabitats and ecological niches into a narrow <lrea.

Vegetation in the study area is predominalllly tundrawilh a groundcover of low-growing plants «I foot high)that includes dW<lrf shrubs, sedges, small herbs, lichens,and mosses. Taller shrubs arc restricted to drainages andsouth facing slopes, Almost the entire coastal plain isclassilied as wetland.

Fivc terrain types predominatc across lhe study area.MOllntain terrain, with its complex and often sparselydistributed vegetation comnltlnities, occurs along thcwuthem periphery. Sedges, tIlssock-fonning sedges, and lowwillow :lnd birch shrubs domin<lte thefoothill terrain. Hillycoastal plains ofgently rolling topography have large areasof pattcrned ground formed by ice-wcdge polygons and frostboils and support tussock tundra, low shrubs, and gr:.ullinoid·dominatcd tundra. Riwr flood plains support localizcd

·habitats of willow thickets as well as a rich diver.;ity of otherplant species and communities, Flat thaw-lake plains ncarthe seacoast have wet and moist graminoid tundra andabundant shallow lakes formed by thawing of pennafrost.

110rc extensive descriptions of the study area can befound in Clough et a1. (1987) and U.S Fish ·and WildlifeService (1982, 19HH).

References

Clough, N, K, P. C. I'auon, and A. C. Chrislensen, editors.1987, ArClk National Wildlife Refugc, Alaska, coast<ll plainresource assessment - report and recommendation to theCongress of the United Slales :md finallcgislali\"cenvin}nrnental impact stat<'lllcnl. U.S. I;ish and WildlifeScrvice, U.S. Geological Sun'ery, :lnd Bureau of LandManagement. W:lshillglOn DC, USA.

G',lrner,G. W., and P. E. Reyno!;!>. editors. 1983. ArcticNational Wildlife Rduge cO:lstal plain resource assessmcnt:19S2 update report, baseline study of the fi,h, wildlife, andtheir habitats. U.S. Fhh and Wildlife Service, Anchor;lge,,\];),ka, US..\.

i\I~LIIC !<.J:/-UUI::: COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU~H,'IARIES 3

km: ~'Wll bou,w,y=-­~Nl"""'nd>ty

J I'<>"rn<rDtw..n..

':=! ~'iI<I""P'''''C'''P1.L'><l'

1002 AreaArctic National Wildlife Refuge

II S 10 :IS 20ml

II a u; 24 32km

Figure 1.2, Geographic map of [he 1002 Area of lhe coastal plain of lhe Arctic National Wildlife Refuge, Alaska.

___. and __, edit{)f~. 19\\4. Arctic N;ltionJI WildlifeRefuge eO;lSl;11 p]:lil1 resource as~e~slllent: 1983 updaterepon. ba~cline sllJdy of the fi~h. wildlik. :lnd their ha]Jil;1lS.U,S. Fi~h and Wildlife Servi.:e. :\nc!1orage, Alask:!, USA.

~,_. ;lIld ..__ ' edillm. 19155. ,\rclic N;ltional \\'ilJlifeRefuge coasl;11 plain resour.:e ;ls~eSsrnenl: 1984 upd;Hereporl, h:lscline ~lUdy of the !i~h, wi!Jlik, :llld llicir hallita!.U.S. fish and Wildlife Sen·ice. AndlOrage, Ala~b, USA.

~__ , and ~_. editors. 1986. Arclic NaliOlnl WildlifeRcruge cnasta] plain resoUTCc a~SC~Sll1ent: !inal repon.ha,e1ine slUdy of the !ish, wildlife, and lheir hahitats. U.S.J:ish and Wildlife Service. ,\nchorJge, Alaska, USA.

~ __, and __. editors. 19R7. Aretic National WildlifeRefuge cnast:!l plain resource a~seS>tllCl1t: 1985 lJpdmcrCjl<)ll, haseline ,tudy of the !ish, wildlife. and lheir halJilat~.

{j .S. rish :UllJ WHdhfc Sen'icc, ,\nchorage, Alaska, US,\.

~kC:lbe, 1'. R.. B. Griffith. N. E. Walsh, D. D. Young, editors,I <)<)2. Re~e;lrch on the p(llenli~1 effects of pctwlc\11l1deveioplllent on wildlirc and Iheir habital. Arclic NationalWildlifc Refu~e inlerim repon. I9SS-1990. AI;I~b fish andWildlife Research Center :l1ld ArClic 1\';llional WildlifeRefuge', fairb;mks, AI;lska. USA.

US fish :md Wildlife Servicc. 1%2. Arctic National WildlifeRefuge C()~st;r1 plain resource ;Issessrncrll: initial rcponb:lseline stud)' of the !ish, wildlife, and their ll:lbil:l1s. U,S.Pi,h ;lnd Wildlife Service. Rcgion7. Anchorage, Abska.USA.

]<)8S. Arclic 1\'alion:l1 Wildlife lkfuge finalcOlllprehensive erJllsen'ation plan, envirunmcnt:tl imp;lclstatemel11. wildeme.<.' revicw. and wild river plan~, U.S. Fishand Wildlife Service, Anchorage, Alaska, USA.

·1 BIOLOGICAL SCIENCE REPORT USGSfBRD 10m-ODOI

Section 2: Land Cover

Janel C. Jorgensoll, PC/a C. Joria, alltl David C.Doug/as

Vegetation Mapping of the Arctic Refuge CoastalPlaIn

Documenting the distribmion of Iand·cover types onthe Arctic National Wildlife Refuge coastal pl:lin is thefOllndation for impact asseSsmenl and mitigation ofpotelllial oil explor:llion and development. Vegetationmaps facilitate wildlife studies by allowing biologists toquantify the availability of important wildlife habitats,investigatc the relationships between animallociltions andthe distribution or juxtaposition of habitat types, andassess or cxtrapolate habitat characteristics acrossregional areas.

To meet the needs of refuge managcrs and biologists,Sil!dlite imagery was chosen as the most cost-effectivemethod for mapping the large, remote landscape of the1002 Area.

Objectives of our study were the following: I)evaluate a vegetation classification scheme for use inmapping; 2) detenninc optimal methods for producing asatellite-oosed vegetation map that adcqu:llcly mct theneeds of the wildlife research and managementobjectives; 3) produce a digit,\1 vegetation map for theArctic Refuge coastal plain using Landsat-lllelllatic!'.Iapper (TM) satellite imagcf}', existing geobotanicalclassifications, ground data, and aerial photographs; and4) pt:rform an accuracy :l.~SeSSlllellt of the map.

The land-c()\'er classification _~cheille developed forthe mapping pmject was based on Walkcr's hierarchicalvegetal ion classification system for northern Alaska(Walker 1983). During the developmellt of the map, thescheme was altered slightly to provide a group of land­Cover classes lhat wcre more cornp,lIible wilh theill fOrmation contelll of the L:llItlS:Il-Tl'\l .~pc-ctral data andancillary data. Wildlife biologists were consulted tocnsure that the system included lantl-cover types relevantto wildlife habitat_~tlldie.~,

We conducted a preliminary assessment of mappingtundra habitats with Landsat-TI'-I and SPOT satelliteimage dala. W... IIsed:m intcc-riltioll of the 2 d;\ta sourcc.~

for one study area and uSl.'d Landsat-Tl'Il e;~c1usivclv foranother. Results indicated that the expense (at the ti'me ofthe study) of illlcgr.1til1g SPOT tlat,1 would not be COSIeffeclive for the enlire mapping project. Landsat-TI\Imethods, however, could improve existing maps madepreviollsly wilh Landsat-1'.ISS data duc to Tr'\l's finerspatial resolution and additional spectral hands,Therefore, further sltldics focused on using the Landsat­1'1'1'1 data.

We eValuated 3 methods for producing a lilnd-coverlllap from Landsat-TM data: I) a supervised classificationapproach where spectral categories \",ere defined byreferencc to field d,1t'1; 2) nn unsupervised nppro:lchwhere spectral Ciltegories were defined by a stlltisticalclustering algorithm without reference to field data; and3) a modeling approach where the unsupervisedclassification was combined with ancillary dilla abollt thelandscape, such as terrain types, slope, and elevation(Jona and Jorgenson 1996). Accuracy asse~'Sl1lents

indicated that modeling was the best approach due to

limited spectral differences among several tundravegctation types.

Spatial data used to produce the land-cover mapincluded 2 Landsat-TI\-I multispectral images, digitalelevation data (including derived slope and sun­illumination themcs), and maps of riparian zones andterrain types (including hilly coastal plains, foothills,mountains, thaw-lake plilins, and floodplains), Each ofthesc data sources comprised athelilatic layer in ageographic information system (GIS).

Field dala were collected ,Il 102 sites in lhe ArcticRefuge, with 5 to 20 plots established in diffcrent hmd.cover types at each site over 4 years. The samplinglOCations were digitized and a GIS thcme of field-verifiedland-cover types was produced.

Field data were cross-referenced with the statisticallygenerated spectral classes to determine the most commonland-cover type associated with each of the spectralclasses. Because many spectral classes represented mOrethan one Iand-covcr type, the ancilillry, non-spectral datalayers were Ilsed to improve the classification (Hutchison1982), Each spcctral class was cross-tabulated with thefield land-cover, terrain type, elevation, sun-illumin>ltlon,and slope layers. 111ese tables were used to guide thcmodeling of decision rules fOr splilling confoundcdspectral e1asses into separatc land,cover types.

The land-cover class assigned to each unit area Oil themap (30-1Il~ pixel) depended on its spectral class andassociated ancillaf}' data, most commonly slope andterrain type. Preliminary land-covcr maps were produced,>lnd then the distribution of each land·cover class WilSviewed in conjullction with color-infrared aerialphotographs showing vegctation to judge the lIlapaccuracy. Additional field data were gathered for problcm~~e:ls, and thc deci.~ioll mles were modified as necessary.1he process was repeated through several itcrationsbefore the final map was produced.

Thc ll1:lpping methods, datll, sUllllllary statistics, andan accuracy assessment were presented in a map user'sguide (Jorgenson et a\. 1994). The ima~e processingmethods were preSellted ill more detail in Joria and~o~genson 1996. Sixteen land-cover classcs werc mappedlFlg. .1.1.). They included: I) wet gralllinoid tundra, 2) wetgraJlllnOid tundra with lO-50t;(, moist indusiolls, 3) moist

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SW.n..·1ARIES 5

Land-Cover TypesCoastal Plain, Arctic NWR

B

-- Arctic NVVR Bound.uy- - 1002 Boundary-- Kaktovik lnupiat Corp. Boundary • ,

•III 15 :lOrnl

IS 24 12m

SIi'~

HERBACEOUS CLASSES SHRUB CLASSES OTHER CLASSES

WG - ,Vet G'"m;nold Tunc:a Poorly (lrillneOtunc,,, ""lh ~e<.llles or Il'asses "r.d lew ~hrut>!i

0' mosses

WG//, -Wet G,aminO'd Tund'" W.th lG",(,·50%mo<st ,ncius;'Onos. often on low..:er.~ered po~~l)Om:.

0' SllanllmOOr

MS"V _ Mo,st Se6;le·"\~~~ow Tund'" ""/;lh 10%·SO',:" wet Inclus,ons. often On low-centeredpolYQCI\S 0' slr"ngmoor

'~i STT _ Moist Shrub--Tussock Tundra ~nd Wate!·track Shrub Tundra Tussock lund'a dom,Mted byerec! ",;1"", and orch. C' 5hrubby Ofa;nagecomplexes 10 IOC\h,'is

ST - ~'.o,st low Shrub Tund'a Upar<l r.!opes "'t~,

"ICC: \{~~:rm aM torch

sp - Mo,st ShiUb TU~d'a on h'Gh-c<!nwICdpoiYllooS, Shru~ P<:'Y(loo'2cd tunma "'tI1

hcrb~ccous typcs ,n polygon Irou~M

m PV - Pa'1li\1;' Vellcta:cd 10%·50'10~cll"m1Cd

8-\ - earlen <:10% \{eg"!~ted

IC- Icc aM snow

:'.'",:,1 WI, - WatOi

• Sl-l-Shadow

~,

MS _ !lo;si Sedge-"WJow Tund,a Mc>~t lund'a""tft ~edges, mosses and ",nows

MSD _ Moest Sedge·Dryas Tundla Mo,s!tund'a \'.'lh sedGes, m<>ss"". and Drya~ Se<.l«chummocks and rro-.t sc:!rs g.'>'c humrr:ockyappc.lr"ncc

n - Mo<st Sedgc·T~sock Tundra MOiO!tunc'" dorr:,n"tc(l by cottongrass iwsocks W.mUndCl5\Ory 01 01...."'1 ~hrut>:. and mO'Ssc5

;C-l I,T - Drya5·G,am,no,d /<Jp-nc Tundra Mois: to crya:p~e tundra "'lh Cwar( shrut>:.. ll'am,nc><:lC, loroo,1,,:l1~ns. and l>.lre ground

RS - Rlps"an Shrub ....'cll-dra'ncd Mer tcrracm;.....tM 10-"" or tall ""~0"N5

DT - Dryas Rr.'er Te'race V,'eil.,jraine(llr\{crtcHaces ""ih Drya5 mat. other d'u.:ul ~hrubs. forooand hchens

Figure 2.1. Land·cover map of the 1002 Area with corresponding vegetation class names, descriptions, and class codes, Arctic NationalWildlife Refuge, Alaska.

;;cdgc-willow tundr:l With lO·5W;}, wei inclusions. 4)moist sedge-willow lundra, 5) moist sedge-Dryas tundra.6) mOist sedgc-tussock IUndra. 7) moist shrub-tussocktundra. 8) moist low-shmb tundra. 9) moist shrub tundra

on high-ccnlcrcd polygons. 10j Dryas-graminoid alpinctundra, II) riparian shrub. 12) Dryas rivcr tcrracc. 13)partially vcgclalcd, 14) barrcn, 15) icc. and J6) Willer.

6 BIOLOGICAL SCIENCE REPORT USGSfBRD 2002-0001

The land-cover classes are described in dctail in themap user's guide, which includes quantitative vegetationcover data, species lists of typical plant communitiesoccurring in each land-cover class, photographs, :llldcross-reference to 7 other classification systems used innorthem Alaska.

An accuracy assessment was performed with anindependenl data set of 318 vegctation plots that were notllsed to lHake the map. The plol.~ were systematicallylocated across the coastal plain and foothills but notacross the mOllntains. Point-by-point overall agreementbetwecn the mapped Imld·cover classes and the tield­assigned classes was 50% (Table 2.1).

Although land-cover types in the classification systemwere distinct, land-cover types in the field occurredacross a continuum. Almost all of the vegctation in themapped area was less than 0.5 meters tall and thestrUctural and floristic differences alllOng related typeswere not great. Subtle transition zones betwccn land­cover types arc Ch:lf<lcteristic of the vegetation of lowarctic tundra. Most errors reported in the accuracyassessment werc between closely relatcd types that weretypically adjacent and interspersed inlhe field.

Approximatdy 86% of the assessment points wereclassified as the correct type or one of the most closelyrelated other types. Agreemcnt is higher whcn simil.1f

•classes arc combined into the fcwcr, more gcneral classes

lypically used in wildlife studies. For example, when themap W,IS combined into 6 or 7 more generalized classesfor ungul:tte habitat studies, ovcr 70~"o agrcement wasobtained. TIle greater initial detail of the I6-class lllapwas prescrvcd, however, because it alJows adaptability toa wider runge of studies.

Proportional occurrences of the vegetation classesacross the entire coastal plain and within various terraintypes were roughly similar between the mapped classesand the independcnt ground-truth data set (Table 2.2),again with the majority of discrepancy arising bctweenc1osely-rclated vcgetation cOllllllunitics.

The finalland-covcr map is available to the public indigital format at hllp:llagdc.usgs.gov. 111e ancillary GISdata layers (topographic data, digitized field data,accuracy assessmcnt point locations, terrain types, andriparian zones) arc archived at the Arctic NationalWildlifc Refuge headquartcrs in Fairbanks, Alaska.

Because the land-cover map and its associatedlandsc:lpc themes have compatible digital fonnats, theycan easily be applied to a variety of future GISapplications. Additional the Illes can easily be incorporatcdas l\\ore resource infom1ation bccomes available, or asnew management or mitigation needs arc identified.

Table 2.1. Contingency lable used to assess lhe accuracy of the land·cover map of the coastal plain of the Arclic National Wildlife Refuge.A!as~a. Table IXlmpares the map's coastal plain and foolhililand-cover classes (rows. ordered by ecological continuum) with field-assignedclasses (columns) from an independent syslemalically·sampled data set of 318 points. Land·cover class codes are defined in Fig. 2.1.

Lao.CoverCloss

WG

WGM

MSW

MS

MSD

TT

STT

Sp

ST

DT

p

v "Agreo

WG 211111 : r iii 11

WGM 7 i 19 i 4 i 4 : 6 I I 1 I i 1 ; 2 iMSW 4 I 9 I '2 I 8 i I I 2' I I ' i iMS I i 4 i 15 4 5! ; I! ! i IMSD I 5 I 4 i 11 18 14 i 1 : j 2 i i 1 i I I

TT 1'1612'851i3i' 1'1 I I

5

i 44

i35! 28

I 56

! 77

40

43

34

54

32

"SIT 1 1 I I 1 1: 5 I 13 i 2 2 I I I I 25 52

SP i1! I : i1! 16 Iii I i8 75

ST il' ilj 2 !! I 450

AT -I---ci--jl--_!i-_-'- +I__if--~,~'-,---.:-'-Lt-::-_-:-+1__:--1_+1__-c2:'-f--.::50~1RS ! i !: 123 1 ill 633

DT iii: i 116! i I 786

PV ! 2 I i ~ 1 I 1! 1 I 1 I i 6 17

SA I'; I !, i I i81 9 89

I-.:W::A':::_1-c-:-+1-c:_i i !: I 1 I I 1 I 4 6 67TOTAL 13 I 40 (-3c,:-ji-4C2:--:3C'----='c'-+i--18:-r-'c5--5:-t--6-+i--=5---'C.,-+i-c'-I 10 i 5 318

~"Agree 15 I 47 I 39 i 36 i 49 67 I 72 40 40 17 i 40 I 43 I 100 I 80 80 50

ARcnc REFUGE COASTAL PLAIN TErUU:STRIAL WILDLIFE RESEARCH SUll,'t~IARIES 7

Table 2.2, Percent of each land-C()ver class in the land-cover map of the coastal plain of the Arctic National Wildlife Refuge, Alaska, and thepercent partitioned among vanous terra'lf\ types, Land-cover dass codes are defined 'If\ Fig. 2.1.

landCoverClass

EntireMap

EnUreCoastal

Plalna

Mountain Foothill HillyCoastal

Plain

ThawlakePlain

Flood­plain

Riparian

Zoneb

9WGM

!--,w"G,,- -2 _,._....£Hl: ,. ~ __<_1J~L. 41~L. .,,'_... ~ "~L ~._

13/91 <:1 1101 21 19\ 23 39120l

MSW 6 9/10\ <:1 4171 10110\ I 23 17113\ 2

MSD

TT

10

---- - ________ ..§lj?.QJ~ 1--- ...1.. ., _..___~_(m_ .._ -.. " - _.t~J~?J 1---

13/12l 3 17/12\ 20151

21 122l <1 32 133) 23'29'

6

8

<1

__________}.J.l.!3L. -' _

"",4 '2'

.. -S..

rS"TT'--_-I~_ .._,~. ..J£{§L .1_ _ ,,~~J!.!L _~ .._~ __~.!JQL ~__.._~11QL . 2_

rS"T__1 -"s-l-__,,3.C,,'1lL+ 8 6121 °'" __ ° <110'

SP 1 1 14\ <:1 1 lSI 1 101 <:1 f 101

AT 2/1\ M 3/4\ Oml 0 <:1~

8PV

rR",S'___1__._-"+ 'C1(~J <:1 1 (Ol , <:1 :01 ! ~1_--",-",I4'L,+----,1"-j8OT 1 1 2/3\ <:1 <:1 ml <:1 ~O\ I <:1 51101 14

6 I 2/21 14 1111 <:1101 I <:1 2161

rB",A ..__--'_~._j ""'" ..?JgL .

[-"'c'-_-t 1

r~"':'---,I---6r---=~-10383523271

M • __2 .. __.._~j§L _-._~

CC!O:L1----',:'-r----'- ~ ::: 1:<:1 r <:110\

32 __________L{9J _____~_L(QL_- -.._------ ,,---- -3 <:1 (0) ____~1 {OJ

<1 <, IOJ 2 (1)

'6 <, '" <:1 101 I7073 6397 1810 I12145Sq-kmd 18501

a Entire map excludmg lho mountain terraIn type.b Riparian zone Is Included within the floodplaIn tanaln typo,c Number in parentheses is tho percent covor for each land-cov!1r typo as estimated by an independent systematic field samplo of 756

points.d Number of squam-kilomaters In each terrain type.

References

Hutchinson, c:. F. 1982. TechniquC's for comhining LanJs:ll andancillary data for digital classification improvement.I'holOgramllletric Engineering :ll1d Remote Sensing,1R{l):123-130.

Jurgenson, J. c., P. E. Jori:l, T. R. McC:lbe, n. R. Reitz, M. K.RaynolJs, 1\1. Emcrs.l\1. t\. \\'illl1~. 1994. Users guide forthe land-cover map of the coastal plain of the ArelieNat'lnnal Wildlife Refugl~. U.S, I'ish and Wildlife Sl'rvice,Fairbanks, Ala"b, USA.

Jorin, P. E., and J. C. Jorgenson. 19%. Comparison of threcmethotls for mapping tundra with LANDSA"I"·TM digitaldata. PholOgr:lmmcnic Engineering and Remote Sensing62(2):163-169.

Walker, D. A. 19S3.:\ hierarchical tundra \'cgetalionclassification cspecially (ksigned for mapping in northernAlasb. Pages 1332-1337 ill PeOllafrtlst: FourthInternational Cnllferl~nce. Proceedings, 17-22 July 19S3,University of AI:lska, F:lirb,lnb, NallOllill Academy Press,Washington DC, USA.

8 BIOLOGICAL SCIENCE REPORT USGS/BRD 2002-0001

I,'IIi!i

Section 3: The Porcupine Caribou Herd

IJrm} Gnffilh, D(wid C. Douglas, Noreen E WaLxh.Donald D. YOIlIlg, Thomas R. McCabe, Donald E.Russell, Robert G. White, Raymolld D. Cameron, {1m!Kenneth R. Whillcn

Documentation of lhe natural range of variation inecological, life history. and physiologic.:11 characteristicsof caribou (Ranglfer /armuhlS) of the Porcupine caribouherd is a necessary base for detecting or predicting an)'potential effects of industrial developmellt on lheperformance (e.g., distribution, demography, weight-gainof individuals) of Ihe herd. To Jcmollslr:llc an effect ofdevelopment, post-development pCrfOnll<1nCC Illust differfrom pre-deVelopment pCrfOfm:lllCC whik accounting forany natural envirolllllcntaltrends.

We had 2 working hypotheses for our investigations:I) performance of lhe Porcupinc caribou herd wasassociated wilh enYlronmental pallems and habitatquality, and 2) acces.~ 10 important habitats was a keyinfluence on demography.

We s()ught to document the range of natural variationin habitat conditions, herd size, demography (definedhere as sllrvival and reproduction). sources [lild magnitudeof mortality, ~listribution, habital usc, and weight gain andloss; and to develop an understanding of the interactionsamong these characterislics of lhe herd.

III addition, we invesligated ways that we could me thisbackground information, combined Wilh auxiliaryinformation from the adjacent Central Arctic caribou herd,\{} predict lhe din..'Clion iUld lllilgnitudc ofany potenti;11 effectsof industrial oil development in the 1002 Area of the ArclicNational Wildlife Refuge on Porcupine caribou herd calfsun'ival on lhe herd's calving grounds during June.

Data, Methods and Assumptions

This work focused on the calving and post-calvingseasons of the Porcupilll:: caribou herd. 111e call'ilJgseason WilS defincd a.~ the 3-week period thaI bcgan wilhthe birth of calves (spring). Post-ell/ring was defined asthe 3-week periodlhat followed the calving Scason (carlysUllllllcr).

Porcupine caribou herd size wa.~ eSlimatcd by lheAlaska Departmellt of Fish and Game (ADF&G) fromaeriOlI pholO-censusl~s during post-calving aggregalions.Only censuses considered reliable by ADF&G were used.Variance in annual censuses due 10 llluitiple observerscounting portions or the photo scts was relatively smallwhL'n COl1lp;lrcu wilh cadi census (+2%) and was iClloredilllhe display I)f annual cell.'ill.'iCS l(;-the ncarest I,O{)Oanimals.

Demngraphy and calf weight-gain were estimaledrrum repeated locations andlor recaplures of radio-

collared animals. Calving distributions were estimatedfwm 767 calving sites of adult (~3 year old) radio­collared female caribou oblained during 1983-2001[average of 40 sites per year; fixed-kernel analyses usingLeast Squares Cross Validation (Silverman 1986. Seamanet a!' 1996. 1998, 1999)1. Concellfrared ca/l'irlg areaswere defined as the annual kemel contour thaI includedcalving sites wilh greater lhan aVerage density (Seaman el;ll. 1998). Annua/ calving grounds were defined as lhe99% kernel utilization distributions Obtailled from annUalcalving sites. ExfeTll of cah'ing was defined as theaggregale e;"lent of all annual calving grounds.

Vegetation types were mapped from Landsat-Thematicl\lapper satellile imagery (Fig. 2.1; Jorgensen el al. 1994)and reduced from 17 to 7 clnsses for caribou habitatanalyses (Fig. 3.1). We estimated the NonmllizedlJijji.'rellce \'egl'fatio!l Index (NDV!) (Tucker 1979,TUcker et al. 19S6) and snowcover from Advanced VeryHigh Resolution Radiometer (AVHRR) data fromNalional Oceanic and Allllospheric Adminislration(NOAA) polar orbiting satellites. Snowcover waseSlimated using a linear regression Ihat we derived bycorrelaling AVHRR infrared reflectance wilh estimates of~'nowcoverextracted from 'lerial photographs collt..'Cted inlhe 1002 Area during lhe snowmelt periods of 1987 and1988 (r =0.87, Tl =80). Cloud contaminaled areas in theAVHRR illlJges were identified (Baglio alld Holroyd1989) ;llld excluded from anal)\~es, as were large waterbodies. AVHRR and 1l1ematic l'.Iapper illlnges weretrnnsfonned 10 an Albers Equal Area projcction and re­s;lmpled to I-km~ pixel sizc.

NDVl indexes the disproportionale refleclance of ncar·infrared radhuion from green vcgetation (Tucker ilndSellars 1986) in the callopy of plant cOllllllunilies. Thus,relationShips between NDVI and tolal green planthiomass or leaf area index (LAI) would be expected 10 bestrongest for planl communities wilh reduced verticaldislributi{)IJ of green biomass ilnd leaf area (e.g.,cOlml1llnities dominated by sedges, grasses, or shortshmbs that arc coml!1on in the Arctic). Due 10 lhe size ofthe pixels (-I klll)) AVHRR data arc linked /l1ore tolandscape processes than 10 individual plant communities(Malingreall and Belward 1992).

Relalively good corre];lliollS have becn oblainedbel ween above ground net primary productivity (ANPP)and se;lsonaJly illlcgrated NDVI (r = 0.89: Pamelo et al.1997), LAI and NDVI when inlegralctl acrossphysiognomic categories (r = 0.97; Shippen et al. 1995),and p!wtosynlhctic biomass and NDVI in small plots (,2 =0.51; Hope d a1. 1993). Because NDVI indexl'd IOtalgreen biomass and carilll)u arc selective feeders (White1983), wc assumed lhal lhe biomass of forages eaten bycaribou was positively correlated Wilh tOla] green biollla~'s

at tIte lall,bcape scale.

AI{(.jIC REFUGE COAS'rAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU/'vl!\IARIES 9

OWet Sedge[J Moist Sedge

111 Herbaceous mm ShrubTussock Tundra Tussock Tundra

CJ Alpine

II Riparian

o Non-Vegetated

Figure 3.1. Land·cover dasses onlhe coastal plain of the Arctic Nalional \Vildl"lfe Refuge, Alaska, and eastward inlo the Yukon Territory, Canada,as generalized for studies of the Porcupine caribou herd. Classes are based on Jorgensen et aI, (1994) as depicted in Fig. 2.1 and are expandedto include Canada usin9 a Canadian Wildlife Service landsat-derived vegetation map of the Northern Yukon. Classes on lhis map and theircorresponding classes in Jorgensen et a1. (1994) include: Wet Graminoid (WG, WGM, some PV), Moist Sedge (MSW, MS, MSD), HerbaceousTussock Tundra (n, SP). Shrub Tussock Tundra (SD), Alpine (ST, AT, some PV), Riparian (RS, DT, some PV), and Non·vegetaled (SA, tC, WA,SH).

\Ve directly estinlated NDVI at 3 times:I) Nf)\'Ccall"ing - composite (Holben 1986) imagesolll'lined a~ dose a'> pos~ihlc to median calving dale~ach year (mean image date of 2 June. 5E = 2.0 days).Snowcover W:1S also estimated from these images.Ncgati\'e NDVI \':llues (areas with snOWc()Ver) wereconvened to 7,ero NDVI.2) NDW~lllid'}IIIIt, - approximately 2 weeks arterc,llving (mean image date of 16 Junc, 5E = 2.6 ,bys).3) NDVljarly·}uly - during the lIrst week of July(Hle;lI1 imag,e dalll of 3 July, 5E = 2.4 days).

From these images \\"e derived 2 additional estimates:I) NDVel'llli' - the pixd-based daily rate of increaseill NJ)VI from calving to mid-June.Z) NDWJ)21 - NDVI on the fixed date of21 Juneeach year (approximately 3 weeks aftcr calving,iinearly interpolated from mid-Junc and e:lrly-Julyimages).

1:1 years whcn Sllowcover was substantial (i.e., 1986,19BB. 1989, 1992, 1997) and NDVI3alving was ncarl,<-'rD, there may have been a small o\'ercstimate ofNDV1_rate, In addition. cloud L:over made it impossiblehl o[)laill a complete image 011 any fixed date. Thus.

NDVI_621 was the most robust NOVl estimate bec;\uSe itwas intcrpolated to a fixed date from 2 snow-free im:lges.

We assumed that NDVl_calving and NDVI~621

represented relative green foragc quantity whileNDVI_rate reflected forage quality because it estimatedthe daily accumulation of new plant tissue which is highlydigestible (Cameron ,lI1d Wl1illcn 1980). The qualityimplication of NDVI_rate was based on the assumptionthat caribou for:lge selectively for the most digestiblefood items (While 1983). Because energy and proteinintake from milk by caribou cal\'cs remains high duringthe first 3 weeb of life and then declines as calvesincrease their intake of vegetation (Whitc and Luid 1984.Parker et al. 1990), we assullled that NDVI~621 estimatedforage availability to lactating females during the 3·weckperiod of peak lactation derl1;lI1d imlllediately afterc.llvillg.

Predator distributions and rdativc densities wereestimated from annual relocations of radio-collaredgrizzly bears (Unus arclOs), 1983-1994, and from aerialsurvey locations of goldcn cagle (Aquila chry.\'(!C{(J.I') neststructures and wolf (Callis IUJ1l1s) dens (Fig. 6.1).

Satellite-collared c:lribou provided supplementalinform;llion on distribution throughom the herd's annu:lIr:lllge. Estimates of minimum daily lllOVemel1\ rates were

J(J IHULUliJ(..,::\L Sl'IENCE REPORT USGSIBRD 2002 .. (}(}()1

obtained frolll satellitc-eollared animals, 1985- I995, andfrom ncar-daily relocations of conventional radio-collaredcalvcs on the calving ground, 1992-1994.

Data were analyzed with contingency tables, linearand stepwise logistic regression, mulli-responsepeTl1lUtation procedures (/'vlRPP, Mielkc and Berry 1982),,lIld analysis ofv:lriance. Abike's Information Criteria(AIC: Akaike 1973. Sakamoto et al. 1986) were used forfinal model selection. Bonferroni procedures were used 10

provide over:11J experiment error protection as'lppropriate. GIS technology, rClllotely-sensed habitatdaw-Iayers, habiwHJemography relationships, andsimulation modeling were uscd to assess pOlential effectsof displacement of c;llving grounds on calf survival eachJune.

Not all types of data were available throughout theelllire primary study period of I(8)-2001. Calf weightsnear birth were estimated from captured 1- and 2-day·oldanimals in 1983-1985, and again in 1992-1994. Calfweight-gains on the calving ground and cow weights inJune and September were estimaled in 1992-1994.

Caribou food habits were estimated during 1973CillOmpson ;lIld 1\kCourt 1(81), 1979.. 1981 (Russell et al.1993), and for this study during 1993-94 frommicrohislOlogical analyses of fecal pellets (Sparks andMalechek 1968) correcled for forag.e digestib il it y(Duquelle 1984).

Annual adull caribou survival was estimated in 1983­1992 (Fancy Cl al. 1994. Walsll et al. 1995). Over-wintercalf survival was estimated in 1983-1985 and 1988(Fancy et al. 1994. Walsh et al. 1995). }IIII£' cal/sun'ivalnhe proportion of parturient wdio-collared femalesretaining live c:llves during the last week of June) W:IS

estim:lted in 1983-1992 (Fancy et al. 1994, Walsh et al.1995) and for this study in 19lJ3-2001.

Calving disnibutions and vegetation types on thecalving grounds were available for all years 1983-2001,hut siltellite-based estimates of NOVI ;llld snowcoverwere only available for the ye;lrs 1985-2001.

The study area covered the annual range of thePorcupine caribou herd (FIg. 3.2), empha.sizing thecalving ground, and was described in the inlroduction to

Figure 3.2. For the Porcupine caribou herd: annual range (wide white solid line), calving sites (yellow points), and aggregate extent of calving(thin solid yellow line), 1983-2001. For the Central Arctic caribou herd: aggregale extenl of calving (thin solid white line) and calving sites (whitepoints), 1980-1995. (Adapted from Wolfe 2000).

'0"'='- ===,.",..",..

ARCrlC REFU(iE COASTAL PLAIN TE1{RJ:STRIAL WILDLIFE RESEARCll StJ!\IMARIES 11

Illis r~pofl anu in th~ 1987 Final Le,gislativcEIl\'ironl1lenlal 11l1P;ICl SIOltcllll'lll (Clough el al. 1%7).

Nutritional Importance of the Calving Ground

Spring arriv;ll Oil thc l..'alving ground is lfll' lime ofminimum bouy reservcs for llilrf/lri('/j( fell/illt'.\' (Ihosc;lhoutto givc birth or accompanicd by vcry young ealn:s)(Chan-MeI,eoul't al. 1999). Thereaftcr, their enagy andprotein requirements rcach thc highest level of the yearduring peak lactation in the first J wcds of June (Whiteand Luick 1984, P;lrker ct al. 1(90). The felllales'appetites are high and forage intake rates can matchbCt;ltiol1 de,mand only when.' primary production is high(White l,'t al. 1975, 1981). Small changcs in nlllritionalcomel]( and digestibility of forage, howevcr, can havcsubstalllial Illultiplier effects on digestible encrgy andprotein illlake (White 1(83), :lJld thus Ill;]y inlluenccllutritional perforl1l:ll1ce of Pllrcupine caribou hcrdfemales on th~ calving grOlHh.!.

Rccent advances ill idell1ifying the basis of selectionof fool! by ungul:nes (k'mollstrate th<!l forage illlake is afunction of ull£lIlate Jllorpholo~y. plant ;lrchitecture, andbiolll;1SS ofaccept;)b!e for:lgc (White et al. 1975. TrmJcll;Ind White II)X I, Spn!in£er ct al. I9SX. Shipley andSpalingcr 1991, Gross et a1. 1993, L:mgv:lln and Hanley1993, Willmhurst alld l:ryxdl 1995). Becall.,e ungulatessdect foragl,' with high dige,tible energy and highdigestible protein (LlIlgvatn and Hanlcy 1993,Willllshurst ;lJ1d Fryxell 19(5), these propl,'11ies arc lhercJe\"alll llle:JSlJre of forage value of habitats at ;IIIY sp;ltialscale (White ct ;11. 1975, White and Trudell 198011.11).Thus. thc forage cllnency for ungulates is primarily ;l

function of digestibiJity of acceptable foods and is not.,imply plant biomass or gros~ energy (Fryxcll 19(1).

The .,ource of protein for fetal growth comes almostl.'xc!usively from !J{)dy protcin of female caribou entcringwilltcr (Gerhart et al. 1(96). Female., with high bodyprotcin in late willieI' produce lhe largest calves (Allaye­Chall 1(91). Early \waning of calves occurs when habitalconditions do not support a protein intake sufficiellt tomeet a minima! rate of body protein depo"ition: milksynthesis then eeascs (Russel! ;lIld White 1(98). Theprotein:cllergy ratio Ill' Ji)rage consunletl during !;Ictationincreases the Illi!k protein intake by calvcs {Chan,Md.cudcl aLl99·J), the lllmt impllrtant milk llutrient affecting c;11fgrowth r<!le at all calf ag('s (White 1(92).

Whell fnr;lgc biomass is low :ll calving, Porcupinel.'arilJoll herd females might bc expectcd to uscmicrohabitats of highcst biollla,s of accept,lble foods antitll seketthe most digcstible foragl,'s from within them. ashas been tlOClIlllCIl!l,'d for c;lribou of the Celllfal Arctic/1I.'rd (Wllite el al. 1(75) :lIld the Weslern Arctic herd(White aud Trudell 19.50!». This ch,lllgc illllle b:1Sis ofselection, from forage biomass to forage digestibilily,

conslitutes seale·d('pendcllt selection (d. Wiens 1989,C)' Neil and King 1998). We pursued this issuc {)f scaledepcndency in hahitat selectlOll by the Porcupille cariboulk'rd at thl' larger sC;lk.s of tlk' allnual calvillg gfOlinds andconcentrated calving areas.

Because the inability to meet laCl;ltion demands maylower the pcrfomwl/({, (i.e., wcighl-gain. survi\':ll) ofcalves, calving ground habitats lllay be iJllPOfI:Ult. TheyIllay be important becau.'e they ean eOlltriblllesubstantially to the female and calf protein bud3cts duringthe calving seaSOll, when maternal pmtcin reserves can below (Gerhart el al. 1996, Chan,l\kLeod et a!. 1999).

Habitat Trends During the Study Period

The climate of thc ,·'retie has heen warming in both~UlJlIller and winter during recent ckcades {Chapman andWal.~h 1993. Groislllan ~t al. 1904, J-loll£hwn et al. 1(95).Temperature increases have been gre;ltcsl ill winter. Tilewarming lws been hetCf{lgeneous al,TOSS the Arctic(Cllapnmn and Wahh 1993, Serrae 20(0), but wasevidclll in spring (Fig. 3.3/1) and wlntcr (Fig. 3.J/J)temper;lIurcs within the Jlorthern par! of the annual rangl,'of lhe Porcupine caribou herd.

An earlier greening ;lIld later Sl~ncsccnee of greenplant biomass in areas Jlorth of 4(}'N (~vIYJleni el al. 1907,1998: Zhou et ;d.10(1) have heell detl,'cted with i\'DVlanti associated with the w;lrrning trend. The earlicrgreening W:IS evi{kllllocaJly withill thc e.\lenl of calving(Fi~, 3.2) of the Porcupine caribou herd in the form of anincreasing relative amount {lfgreen plant bioJllass on21June (NDV IJl21. /~ ::: 0.50. !'::: 0.0(2) dming 1985-/999lFig.3.4),

A \·('ry low value for NDVIJl21 was observt:d in1992, lhe yC;lr that stratospheric ;lerosols from the 19() Ierujltion of l'Ilounl Pinatubo in the Philippines r('ached theArctic in .'pring (~lillnis et;ll. 1993). Both 20()1 and 2000were substantial outliers (RStudent::: ,2..19, -1.86,respecti\'elyJ frOJll tIle relatiollship betwcen NDVI_621and year, 19'<\5'1999 (Fig. 3.4), Bot1l2001 ;md 2000 hadI.'xceptionaJJy 1;ltc sprill£s wilh high .'no\\"cowr al calving.We do (11)[ yet kilO\\' ifll!esc l)utliers indic'ate a change inthe trend ohserwd during 1085-/999.

The :tn·ric Oxci!!arioll (r:ig. 3.5) i.\ celllered over thehigh Arctic and is nne of a nUlll!ler of correJ;lIed indicesof largc'.,cak' atmospheric pressmc differell1iah (e.g.,North ,\tl:lIltic Oscillation, Northern l-Iclllispherie :\nnular!\lode) (Thompson and W;IJl:lce 199B, 20(1). The Aretil,·o.'cillation is the heighl of the le\"el of one,halfat!llosphere of pre.\surc- above thc surface of thl~ l,'anh andis weakly c\)rrel,lled with sllrf:lel,' lClllper:llllrCS(Thompson and W;llJacc 19911). The Arctic O~citr;lIion h,l.\a W;lrm po.,iti\·l' phase when surface prt:;;sur('s are low:lnd warm North :\ll:ulIic water enters the Arclie Oc(':m,

12 BIOLOGICAL SCIENCE REPORT USGSII3RD 2002·(){}()!

,-~-'_.-------c-------.PoY.r,'" Pha><l c( AO

, .~ ,. I l ,< • '\ I

il.!::j o·

"ro-g ·1 .

"If) .21 V POO phJ50 shilb'

.J L~~"3::~.~~.:-~:,:~-_~j"__._,_~_-.~.~i__ ,. y.__•__~I&~5 1(165 \97~ 1%5 1995 2005

Year

Figure 3.5. Standardized values of the Arctic Oscillation (AO) forwinter (January, February, March) and population size of the Porcupinecaribou herd, 1958-2001. Mean value indicated by solid horizontal line.• PDO 'IS the Pacific Decadal Oscillation (Hare and Matuna, 2000).

and a coolnegalivc phase when surface pressures arerelatively high.

Initiation of increasing and decreasing trends in theArctic Oscillation has been coincidelll with phase shifts inthe Pacific Decadal Oscillation in 1977 :md 1989 (Hareand l'.fatuna, 2000) (Fig. 3.5). Correlations between theclosely rehllcd North Atlantic Oscillation and a number ofvegetative and ungulate population characteristics havebeen reported for Northern Europe (Post et al. 1997, Postand Stenseth 1999).

Median annual NOVI at calving (NOVCcalving)within the extent of calving of the Porcupine caribou herdwas positively correlated with the Arctic Oscillation fromthe winter (january, February, l\-'larch) of the previoust"aJclldar year (-15 month lag, r = 0.32, P = 0.011) (Fig.3.6). 'nlis suggested that early forage availability forlactating females was influenced by weather patterns on ahemispheric scale.

Further, the suspected phase .~hift in the ArcticOscillation at the end of the 19HOs (Fig. 3.5) wascoincident with an increase in the frequency of dailytellljJerature excursions above freezing in both the spring(Fig. 3.711) and fall (Fig. 3.7h) on the transitional rangesof the Porcupine caribou herd during the 1990s. There hnsbccn a decre-:lse in the depth and extent of snowcover inNorthwestern Canada ncar the willlering ground.~ of thePorcupine caribou herd during this latter period as well(Brown and Braaten 1998).

'I1ms, forage biomass during peak lactation demandlNOVI~621) increased during the period of study, 1985­1999 (Fig. 3.4), and this positive trend was coincidentwith snlllmer warming on the calving ground (Fig. 3.3a).In addition, forage availability at calving (NDVI_calving)has been positively cl}ffe]aled Wilh hemispheric-scale

.......•-------_•...,.'----~--'c

~ 0:;0 'j' .' ~ ••

m

~ 1ro ,U o~o '\

I,

N ,.5 o?O-1> ,oz

o10 -;~!:,; 0'-;-r07 '-i-~~;_~i;~1 ',';;:;"3T ',,,ch"--;'l:.{"';r&3 - fOOlYear

•

Figure 3.3. Mean temperatures for 2 stations within the Porcupinecaribou herd's aggregate extent of calving (Komakuk Beach andShingle Poinl, Yukon Terrilory, Canada) and 1stalion \\ilhin its winlerrange (Old Crow, Yukon Territory) for al June, and b) winter (January,February, March), 1950-1995.

Figure 3.4. Median Normalized Difference Vegetation Index (NOVI) on21 June wilhin the aggregate extent of calving for the Porcupinecaribou herd, 1983-2001. Values for 2000 and 2001 were outliers(RStudent =-2A9, -2.86, respectively) and excluded from thedisplayed regression line, r ,= OA96, P=0.002.

ji'i,iIiIi,,Ii~iIi

ii

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU1\l!\IARIES 13

-j-I-

I

•

..- -- _.-Decrease Phase

•

•••

•

••

•

IncreJ~C PhJSC

" - Increase - 1970-1988 Decline - 1989-1998Population Trend

tncrease 1970-1989 DeCline - 1990-1999Population Trend

cooI. 16

~o

"f~ 8~

10

.,~•0 ,

coo!\ 16

~o"'§ ,.;

~•,-~.~

3oW~•o

,j

bj

Figure 3.7. Frequency 01 days wilh daytime tempera lures abol'cfreezing 'rn c1) spring (21 March - 30 April) and b) fall (21 September­20 October) on Iransitionat ranges of the Porcupine caribou herdduring lhe herd increase phase, 1970-1988. and the herd decreasephase, 1989·1 99a. Brackets indicate 95% con~dence intervals onmean values.

Figure 3.B. Population size of Ihe Porcupine caribou herd. 1972·2001,estimaled from aerial photo-censuses by the Alaska Depar1ment ofFish and Game.

00.

"'.

<'".

,)C'..':000

;:; D:'~

~ O:illoogo,';'U

GO;:>;)

~onc,~I 0 TO

6 Of.~

Z

Figure 3.6. Mediarl Normalized Difference Vegetalion Irldex at calving(NDVI_calving) wilhin the aggregale extent of calving (Ee) of thePorcupine caribou herd for the currenl year, and winter ArcticOscillation index (AO, January. February, March) for the previouscalendar rear, 1985-2001,

The growth CLlI"\'~ of the Porcllpin~ caribou lll'rdsuggtslcd an approxima1~ 30- to -'to-year cycle ,IfinCfl'ase and decrease in ahundance (Fig. 3.8), Tht herdIlulllhereu-IOO.OOO ill 1972. incrcased:l! about .l.l)'} perycar from 1{]79lhrough 1989 when it reached -In.noO:Ulimals. lhen declined at aboul 3.6(} per y~ar from 19l'\{]tn llJ98 (Fig. 3))). The decline from 1998 to 2001 wasonlyabollt 1.5',:, per yt'ar, aud the Ill'rd no\\' totals-In.on :lIlilllals.lfthe current dtdin~ continues.lheherd \\"ould be expected to again reach lhe lo\\"\.'.';t levelsevcr record\.'d dUfing 200)"2010. If the herd continues l(l

(Ieclinc below -100.000 animals, thcn tlle lenglh of acomplelc Ill'rd cyde flIay cxcccd 30 Yl';JfS.

Herd Dynamics and Demography

·1', ·'0 ·DO 00 O~ 10 T5 ~O ;,~ 30

AO Index (JFM) - Previous Year

atlllospl1l'ric L'onditiolls (fig. 3.6). Courlterilctlng tht.positive trend in forage abundance during peak bct:lliollhas htcn a tendcncy IOwaI'd Illore fr~~ze-tha\\' cyclcs onspring and fall tr:lJ1~ilional ral1g~~ of thc Porcupincc<Jriholl h~rd (Fig. 3.7d,b) coincidcllt with a suspectcdphase shift in the Arctic Oscillation.

These freeze-thaw cyd~~ Ol! Ir:lnsiliOllal and winterranges mily have inlluellced .\now properties. rcduccd acce%10 forage. increased Irave! co.';t,.;, and/or deueased lhe ability(If carihou 10 cscape their pr-:Jmors. Th-:s~ climate­inlluenccu c()I1ditions on transitional/winter r:lngl:s m:ly havecontrihUlnllO thl: decline in size of the PorL'lJpin~ cariboulwrd (Fig. 3.5) in spite or f<lvor;tbl t conditions on the cal vingground. Local and largc"scale cIimale patterns as well ascat:lstrophic events in Ihc Southern Hemisphere (e.g ..eruption of 1\1011rH Pinatubo) apparently hav\.' had majorintlucncc.\ on Porcupinc caribou herd habita1s uuring lheperiod of study anu h;lve set the st:tgc for all o!lscrvationsof Porcupinc caribou llcru distribution :Ind dcm(lgraphkProCl'.~SCS during the past 2 uecades.

14 BIOLOGICAL SCIENCE REPORT USGSiBRD 2002-000J

Porcupine c3ribou herd size appeared correlated withArctic Oscillation nlthough there were too few data 10conduct a proper time series ilnalysis (Fig. 3.5). Incontrast to the Porcupine caribou herd, DIller Alaskabarren-ground caribou herds (Westem Arctic, TcshckpukLafrc, Central Arctic), generally continued 10 increaseduring the downward trend ill the Arctic Oscillatiolllhatwas evident during lhe 1990" (Fig. 3.5),

Capacilyfor growth (defined as tlIe Illaximum realizedlung-Icnn growth rate) of the Porcupine cariboll herdappeared subslalllially less than for oIlIer Alaska herds.Capacity for gn)wth ,\lllong herds of dranwticallydifferent sizes is best visualized by plolting rc!atj\'c herdsi;tes (Fig. 3.9). r.'!aximulll long-term growth rate (-4.9 t.'O,assumed linear, 1979-1989) (Fig. 3.8) of the Porcupinecaribou hcrd was nevcr morc than about half tlle rateobserved for other Alaska barrcn-grollnd caribou herds[Westcrn Arctic herd (1976-1996. -9.5%), TeshekpukLake herd (1978-1993, -13%), Central Arctic herd (1978­1992, -10.3%)1 (Fig. 3.9).

The Porcupine caribou herd was the first AI:lskabarren-ground c:lribou herd co begin and maintain aprolonged decline in the last 2 dec:ldes (Fig. 3.9). Annualsurvival of Porcupine caribou herd adult females was onlyabout 34% (Fancy et al. 1994, Walsh et al. 1995), whichW:IS lower than lhat generally observed in other c:\ribouherds (BergenJd 19S0);and adult female survival mayhave been respollsible for the relatively low growth rateof the Porcupine caribou herd.

Annual calf survival averaged ahOllt 48'7;' with abouthalf (56'70 of the annual 1lI0nalil)' occurring on thecalving ground (Whitten et:ll. 1992, Fancy et al. 1994,Walsh et al. JlJ95).

There werc no significant differcnces inmcanpanurition, calf survival during Junc, or nel nllfproduction (defined as the produc{ (l( panurition rate andJune calf survival) (Fig. 3. lOa-c) bctwcenthc increase.and decreasc phases of the herd (Fig. 3.8). Panurition rateaver.lged 0.81 (range 0.71·0.92) during 1983·2001 (Fig.

,j10 T

,,1 ] fl njt:: II] I) r 11j,i1 h1i:: mIIIIIIIAII!IIIUIII'11 W

1saJ 1$55 smi ir.gi 1993 1995 iwi' ;~ l~;Year

bj ,,

-------_ .•---- ...-.__ ..

PCH -178K.-WAH - 463K

.-.-CAH·27K

--,]j-

TLH - 28K

oj~ 10-f

~ 091-rn ,cc ,

"1?: 0 B -j- j-. '11 ('ro I " 'j'll~""I ,I f1 II 'I'""r' j'" 'I1f);1 ')J ;" !J ;.'c'" i ,j' '1"-1,1" li,,!1 Ii/,'o e" "I I' 1'1'11'"" ii'" ",';"" _",I n Iii5"' ,,1'1,- IIIJi, 111'11-; I' I" ''''I' I, 'I [1't ",' II '1'1' "'11 11' '1' I 'ro 1i"'I'" I ' 1'1., 'i'I, .. ,,;,I,] i "IIa. I·,·· "., Co<: ,. f'.· ," ...• I,' ''-; ". "., I,' "C' ,-. f" '.' .'., '.•O~ ".",.j".~,__....,l.•• j ..",L,,,J,L.L_._..,•. l, ....,.c.l,...,...,L._..,

li'53 I%~ lM7 1,-69 1l.'S1 19$3 I¥.h 19")7 '.'3"-" X>:ll

Year

I

FIgure 3.9. Relative post·calving herd sizes (minimum observed =1.0) of the 4 Alas~a barren·ground caribou herds (PCH =Porcupinecaribou herd; WAH:: Western Arclic herd; CAH :: Cenlral Arctic herd;TlH =Teshekpuk lake herd), 1976-2001. Maximum observedpopulalion size for each herd is noled in [he legend.

Figure 3.10. ReproduClivEl eslimates for [he Porcupine caribou herd,1983-2001: a) parturition rale of adull females, b) calf survival frombirth through Ihe last week of June, and c} net calfproduction (theproduct or parturition rate and calf survival].

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMl'I·1ARIES 15

3.lOa) and did not differ between the increase plH'lse(0.80, SE =0.04, 1983-1989) and the decrease phase(0.82. SE = 0.08.1990-2001).

Calf survival during June was quite high and aver;lged0.75 (range 0.57-0.94) during 1983-2001 (Fig. 3. lOb) butdid not differ between the. increase phase (0.71, SE =(J.07, 1983-1989) and the decrease phase (0.79, SE =0.13, 1990-2001). Net calf production averaged 0.62during 1983-2001 (range 0.50-0.82) (Fig. 3.lOc) and did110t differ between the increase phase (0.58, SE = 0.06,1983-1989) and the decrease phase (0.63, SE::: 0.13,1990-2001). For all these demographic characteristics,variance tended to be greater during the decrease thanduring the increase phase of the herd.

Because average parturition, calf survival during June,and net calf production did not differ between theincrease and decrease phases of the Porcupine caribouherd, 1983·2001, a reduction in adult, sub-adult, and/orcalf survival while animals were off the calving ground inlate-summer through winter must have accompanied theherd decline. Emigration to the adjacent Central Arcticherd was an unlikely cause of the Porcupine caribou herddecline bec3use satellite-collared animals thatoccasionally (4 OUI of 167 coIl3r-years) wintered with theCentr31 Arctic herd, returned to the Porcupine caribouherd the following summer.

Periodic lows in net calf production ami calf survivalduring June 0992,1993,1997; Figs. 3. lOb, c) were no!sufficient to maintain the herd declinc (5. A. Arthur,Alaska Department of Fish 3nd Game, personalcOllllllunication). LJnfortu1l3tely, a complete record of'ldult, sub-adult, and calf survival estimates was not:lv3ilable for late-summer through winter during thedecrease phase of the herd, 1989-2001.

Seasonal Distribution and Movements

The Porcupine caribou herd caribou winterl:d (15Novemba ~ 1·1 April) in Alaska south of the BrooksHallge and in Callad3 in tlle Richardson and Ogilvie~'lountalns in the Yukon Territory (Fig. 3.11). Theirannual range enwll1passcd -290,000 klll" (Fig. 3.2). lllCe;o;tent of calving encl)l1lpassed -36,(}{)0 km". Springmigration to the anlluOlI calving grounds began in mid­April and continued through April and ~lay (Fig. 3.11).Return to fall/winter ranges began with departure fromthe annual calving grounds in late-June and early-July(Fig. 3.11). In fall (15 Scptember-14 Nm·ember). thePorcupine cnriboll herd was distributed widely.

r...lininlllm daily travel rates of parturient females werevariable throughout the year (Fig. 3.12). Non-parturientfemales had similar movemcnt rates. ~linirnull1 movementoccurred during winter. I\fovement began increasing inmid-April with initiation of migrmion 10 the annual

c31ving ground and was directional toward the annualcalving ground.

After their calves were born, the direction ofmovement of satellite-collared parturient females wasrandom for 20 days (Fancy and Whinen 1991). CalflHovement rate (minimum, stmightline, estimated fromconventional radio-collars) in the ye3rs 1992-1994 wasabout 2.5 km/day during the first week after birth. Thef3te incrcased gradually during the next wcek to nbout 5kill/day and then inere;lsed through the cnd of June 10approxinl3tely 15·20 km/day. As fellwles 3nd calvesdeparted the c31ving ground in late June and early July,some individual calves traveled as much as 90 km/day.Helativcly high rate of movement continued throughoutJuly. Because movement rates were low during thecalving season and directioll of movement was randomfor 20 days after birth (Fancy and Whitten 1(91), thedistribution of c<llving sites was assumed to berepresentative of habiwt usc by caribou through 21 Jtlne.

Movement declined during August perhaps inresponse to harassment by Oestrid flies or to localizedforagc abundance. 1'.fovelllcnt increased during the pre-mtperiod in late-September and Octobt.:r and then rcached aminimum again by mid-November. TIle average female ofthe Porcupine caribou herd traveled approximately 4,355km aillUlally (Fancy et al. 1989).

During 1985-1992. median arrival of satellite-collaredparturient females on the allnual c31ving ground rangedfrom 17 MayA Junc and median date of departure rangedfrom 3-26 July. Non-parturient females tended to lagslightly behind and south of the parturient females fromearly-May through calving (Whitten et 31. 1(92), blltwithin I week aftcr calving, paI1mient :Illd nOli-parturientfemale distributions were essentially coincident.

Lcngth of stay on the annual calving ground rangedfrom 34-67 days. Caribou have tended to depart theannual calving grounds earlier since 1995 (F. J. Maller,U.S. Fish and Wildlife Service, personal comlllulliemion).This trend lllay have been relnted to Illore advanced plantphcnology within the e;o;tent of calving in late June duringthe late 1990s (Fig. 3.4).

!',.·ledian calving date, 1983-1996. was I June (range 30May-6 June) with 50S;' of ;Iflllual calving occurring within2 days of the annualmedlan calving dale. No temporaltrends were cvident in median calving date, and annualcalf survival was not related to mcdian calving date (P >0.1)5).

Sizes and locations of annual calving distributionswere quite variable. Annual calving grounds encompassed.1,(j72-1(j,667 kl11~ during 1983-2001 (Fig. 3.13, TableJ.l). Similar distributions were observed during acrialsurveys. 1972-1982 (Figs. 11-5 ill Clough et ai. 1(87). Onaverage, concentratcd c<llving area.~ occupied 12.3'.';'(range 0.7-25'7(0) of the annllal (alving grounds (255-

16 BIOLOGICAL SCIENCE REPORT USGSIBRD 2002·(){){)[

Figure 3.11. Distribution of satellite-collared female caribou of the Porcupine caribou herd during 7 time periods,1985·1995. An average of 10 animals (range 4-17) were collared each year yielding 14,447 observations; 87% ofthese Observations were obtained 1985-1990. Not included were the locations 013 females that each spent onewinler with the adjacent Cenlral Arctic herd.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUl\llvtARIES 17

2,548 km2) and contained 471,:;' (range 29-6) %) of calvinglocations.

There was no concentrated calving area in 2001 whenthe spring was very late and the eXlent of calving wasalmost completely snow covered. Density of pal1urientfemales in the concentrated calving area rangedapproximately 13-106/km' over the years and averaged 7limes (range 3.7-10.8) higher than outside theconcentrated calving area each year (Table 3.1). None oflhese estimates differed between the increase anddecrease phases of the herd (P > 0.05). Since 1972, therehave been only 2 ye::lfS (2000, 2001) when all calvingoccurred in Canada and I addition::!1 year (1982) when allconcentrated cal\'ing occurred in Canada.

Neither the areas of annual calving grounds nor areasof concentrated calving areas were correlated (P > 0.05)with the number of calving sites, with the estimMednumber of parturient females ill the herd, with the ~rcelll

of the extent of calving that was snow free, or with anygreenness (NDVI) estimate in either the extent of calvingor the annual calving grounds. Thus, neither herd size norhabit;lI charactcristic.~ were clearly related to calvingground si7.e. Factors affecting calving ground sil.c remainunclear.

Distribution of calving sites differed (MRPP, P < 0.05)among all sllccessive years, 1983-2001, except 1983-1984when the number of calving sites obtained from radio­collared females was lowest and 2000-2001 when latesprings restricted calving to Canada (Table 3.1). TIlerewas no uni·directional trend to shifts in location of annualcalving grounds or cOllcentnlled calving areas (Rayleigh'sTcst, P = 0.870 and 0.740, r~spectively). During 1983~

1994, panurient females displayed no among-year fidelityto the concentrated calving area (P = 0.951) nor anyhabitat attribute for calving (P > 0.135), bllt females thatcalved in the 1002 Area returned there for calving in thefollowing year morc often than expected (P =0.024).

The percent of femalcs calving in the 1002 Area in theyears 1983-2001 was quite variable, averaging 43%(range 0-92%) bUI not differing (P =0.128) bctween thedecre,lse (50%, SE = 32%) and the increase phase (JO%,SE = 23%) of the herd (Fig. 3.14). The proportion of theconcentrated calving area tlwt was in the 1002 Arl~a

followed a similar trend. As the relativc amount of grccnbiomass at calving within the extcnt of calving(NDVI_calving) increased beCall.~C of earlicr springs. thcpercent of fcmales calving in the 1002 Arca increased (r

= 0.68, P < 0.001) (Fig. 3.15). Tlms. the averagcproponion of Porcupinc caribou herd females thai calvein the 1002 Area lllay increase if thc climate conti niles 10

warm.'nlC general location of calving in the years 1983-200j

was related to the wintcr Arctic Oscillalion (January,Fcbmary, tvl ",lrch) during previolls calendar year,aJ1pr(),~illlately 15 months bdore calving. In years when

Jun Jul Au,) S<Ip OCt I~,,", Doc

Dale

Flguro 3.12, Minimum median daily movement rate of panurientsatellite-colJared females of the Porcupine caribou herd, 1985-1995.Values calcula1ed from no more than one location per day. An averageof 10 animats (range 4-17) were collared each year yielding 14,447observations; 87% of lhese observations were obtained 1985-1990.Not included are the data for 3 females that each spent one winter ....ilhthe adjacent Central Arctic herd.

thc Arctic Oscillation was positive, more than half of theconcentrated calving area was likely 10 be located on theAla.~ka portion of the coasl;ll plain (83.3% of the years,Fisher's Exact Test, P = 0.045). Similarly, there was atendency (66.7% of years, Fisher's Exact Test, P = 0.057)for more than half the females to calve in the 1002 Areawhen Ille Arctic Oscillation in the previous calendarwinter was positive.

TI1C time delay in correlation between the ArcticOscillation and calving location and between Ihe ArcticOscillation and NDVLcalving (Fig, 3.6) Illay have beenrelated to a I-year delay between tiller fonllation andflowcr production for EriopllOrum l'agil1l1twn(collongrass) (Billings and Mooney 1968, Bliss 1971).Imlllature cotlongrass flowers havc been a dominant fooditcm for Porcupine caribou herd when they have calvedoillhe Arctic Refuge coastal plain. Cottollgrass tillerformation is prob;lbly related to the availability ofresources (moisl\lre and soil nlllrienls).

Positive phases of the Arctic Oscillation lila)' havcenhanced rcsnurce availability, incrCilsed tiller productionin the previolls year, and resulted in increased flowerproduction during the CUlTent spring. We would expectIhat thc incre;l.~cd greenness at calving (NDV13alving)might rellect le<lf ;\rea of cotlollgrass tillers, rather thanthe pale grecn immature llowers.

During post-calving (>3 weeks after calf binh).Porcupine herd cariboll (reganlless of calving location)tended to move westward (fig. 3.11). Even in exceptionalyears when calving Ol:currec! far to the e;lst in Canada(e.g., 2000, 2(01) (Fig. 3.13) caribou reached lhe ArcticRefuge coastal plain and portions of the 1002 Area bylate-June or July (S. :\. Arthur...\laska DL'panmellt of Fish

UIULUlill.'AL SCIENCE REPORT USGSIBRD 2002-0001

Figure 3.13, Calving dislributions of the Porcupine caribou herd, 1983-2001, as eslimated from fixed kernel analyses of lhe sites where radio·collared femates were firsl observed with calves during repeated aerial surveys in May and June, There are 3 zones: 1) concentraled calving area(shO'Nn in dark gray), the conlour enclosing calVing sitos with greater lhan average fixed kernel densily, 2) annual CiJlving ground (medium gray),the 99% fixed kernel ulilizalion dislribution for ayear, and 3) aggregate exlent of calving (light gray), the ouler perimeter of all annual calvinggrounds, No concenlrated calving was detecled in 2001.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUl\IMARIES 19

Tabla 3.1. Number of calving sites, number of calving sites in the concentrated calving area (CCA), area (km1) of CCA, area (km l) of annual

calving ground (ACG), ralio of sizes of CCA to ACG, population size of the Porcupine caribou herd, percent of radio-collared female caribou thatcalved in the CCA, percent of radio-collared female caribou that calved in the 1002 Area, percent of the CCA within the 1002 Area, and percent ofthe ACG wilhin the 1002 Area, 1983-2001, Alaska, USA, and Yukon Terrilory, Canada.

CalVing Sitos In CCA ACG Ralio Population %females %Iemales %CCA %ACGYear SHes CCA Area Area CCA/ACG Sizo (K) InCCA In 1002 In 1002 In 1002

1983 18 11 2,584 10,064 0.25 135 55.6 61.1 62.4 42.8

1984 18 11 839 6,599 0.13 61.1 33.3 19.8 39.2

1985 34 16 1,585 10,784 0.15 47.1 55.9 69.2 36.8

1986 20 8 419 5,432 0.08 40.0 10.0 28.8 8.4

1987 36 15 479 6,048 0.08 165 44.4 13.9 14.2 15.7

1988 61 24 267 3,823 0.07 39.3 1.6 0.0 5.9

1989 51 15 255 3,872 0.07 178 29.4 33.3 59.3 30.1

1900 53 22 1,167 8,379 0.14 39.6 69.8 100.0 47.2

1991 43 21 731 5,767 0.13 48.8 88.4 92.5 68.6

1992 43 18 2.174 16,667 0.13 157 41.9 41.9 79.1 22.5

1993 35 18 1,401 9,098 0.15 51.4 57.1 70.2 40.3

1994 79 33 814 6,602 0.12 152 41.8 84.6 n.3 54.8

1995 60 31 827 5,141 0.16 51.7 91.7 100.0 71.2

1996 65 30 1,354 9,453 0.14 46.2 53.8 90.6 33.•1997 29 15 530 5.661 0.09 51.7 31.0 33.7 31.8

1998 39 20 789 6,316 0.12 128 51.3 84.6 93.4 73.1

1999 20 • 601 7,820 0.08 45.0 20.0 9.3 30.42000 22 13 791 8,541 0.12 59.1 0.0 0.0 0.0

2001 41 a 10,602 123 0.0 0.0

average 40 18 976 7,604 0.12 148 47.0 42.7 55.5 34.3minimum 18 8 255 3,872 0.07 123 29.4 0.0 0.0 0.0

maximum 79 33 2,548 16,667 0.25 178 61.1 91.7 100.0 73.1-

SE 18 7 630 3,060 0.04 20 7.8 30.1 35.9 22.5

a No concentraled calving was detected in 2001 .

0.00 0,05 0,10 0.15 0.20 025 0,30NDVLcalving - Extent of Calving

0.35

Figure 3.14. Percent of radio·collared Porcupine caribou herd femalesthat calved in the 1002 Area of the Arctic National Wildlife Refuge,Alaska, 1983-2001.

Figure 3.15. Pelcent of radio·collared Porcupine caribou herd femalesthai calved withill the 1002 Area of the Arctic National Wildlife Refuge,Alaska, in relalion to the median Normalized Difference VegelationIndex at catvillg (NDVI_calving) ',~i!hillthe aggregate eXlent of calving,1985·2001. Point legends indicale tM year of the estimates.

20 BIOLOGICAL SCIENCE REPORT USGSfBRD 2002-0001

and Game, personal communication). As a result of thesewestward movements, essentially lhe entire 1002 Areawas eventually used by late June or early July. ~Iost ofthe usc of the weslernmost portion of the 1002 Area bysatellite-collared females of the Porcupine caribou herdoccurred during 24 Jllne-14 August (Fig. 3. I I).

Foraging on the Calving Ground

The calving season diet of Porcupine herd c;lribouduring 1993-1994, when concentrated calving wasprimarily in the 1002 Area (Fig. 3.13), W;lS dominated(76-82''J-o) by immature flowers of cOllongrass from Ihetime the caribou arrived on the calving ground until about16-18 June (Figs. 3.16a, 3.170). Similar diets wereobserved in 1973 (Thompson and l",kCOllrt 1981), but the]oc<llion of concentrated calving in that ycar was notdocumcnted (Clough ct al. 1987).

Diet was relatively consistent betwcen years, butsomewhat more variable in 1994, and not related toavcmgc daily weight-gain of calves in 1993 and 1994.Both couongrass flowers and young willow (Salix spp.)leaves arc easily digestible and arc COllllllon forage ofupland calving caribou when they arc aV<lilable (e.g.,

-)

~

i~jlUJ~.W,; "'i,L'~i"ll'~ ~ ~ ~" ~- JVHEl9"34 JULY

CD o:,ltOl'J{Jf<l~~ rnm 'N,~ja" [-J liche~~ • HClbs

b)

.....~ Coltongrass - ...~ Willow

Figure 3.16. Porcupine caribou herd <1) diel composition and b)median phenology 01 major forage ilems. 1993. Diel compositionestimated from microhistological analj'sis of lecal pellets, corrected fordigestibility. Phenology scores lor collongrass: 1=leaves only, 2 =flowers in boot. 3 =early flower, 4=full flo'....er; and for willow: 1=dormant, 2:: bud swelling, 3 =ieaf unfOlding, 4:: fuillenf.

_____-,- ,··__' __T ' , __,_"_" __'_'"

Thompson and l\kCourt 19H I, Kurop:!1 1984, Russell ctal.1993). Cotlongrass flowers wcre most common in theveget<llion type herhaccolls tllssock tundra, and willowwas most common in shrub tussock tundra and rip<lTianshnlb vegel<ltion types (Jorgenscn ct al. 1994).Herbaceous planl.~ were ubiquitous.

Dictary ~hins within the 1993 and 1994 c<llvingseaSO/1S apparently allowed caribou 10 increa~e nutrientconcentration in their diet <IS the season progressed. Bymid-JUlie, 1993-1994, as coUongrass nowers matured, thelcu\'es of willows unfolded (Figs. 3.16h, 3.17[,). lllen,within about·1 day.~ (Fig.~. 3.1611, 3.17(1), caribou dietshifted to an apprOXimate 50:50 mix of willow andherbaccl)us plants.

The dict ~hift resulted in an increase of dietarynitrogen concentration (from Yi; \l) 401:) and a decrease inNelllr:ll Detergcnt Fiber (NDF) conccntration (from 57%to 2iS,) b:lsed 011 nUlritiollal analy~cs of coltnngrass andwillow of appropri:lte phenological ~tages frolll thecalving ground. Availahk biollla~sof willow likdyexceeded the billl1la~~ of cot!()ngrass tlowers during thediet shift and tllereafter.

Figure 3.17. Porcupine caribou herd a) diet composition and b)median phenology of major forage items, 1994. Diet compositionestimated from microhis!olOjJical analysis of fecal pellets, corrected fordigestibility. Phenology scores for cottongrass: 1 =leaves only, 2 =nowers in boot, 3 =early flower, 4=lull nower; and, for willow: 1 =dormant 2 = bud swelling. 3 = leaf unfolding, 4 =fuli leaf.

·-"-"-r·-j-·-·, .\!i

'"I

......... Willow

. .-..'\ II \ ,I! 1/', ..

._._._..... ·.·.·ft,

'LV> '"-J", ,.\--'"D.lIC·1Ga1

Diet Shift - ..I!

~_. Coltongrass

;',•

. . .--1'\ 1\ /

\.1 ".ro,.,j ft • ft

b)

I ARCTIC REfUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU/\.IMARrES 21

Caribou maintained the willow and herbaceous dietuntil they depaned the calving ground near the cnd ofJunc. Because climate warming and earlier gn::cning Illayincrease the carbon:nilIogcn ratios of individual foragespecies and reduce their quality on fixed dates (Walsh CI

al. 1997), rapid shifting among forage species lllay allowcaribou 10 accommodate lime-specific reduction illnutritional quality of individual plant species thataccompanies climate warming.

Diet of Porcupine herd caribou was substantiallydifferent when they used the Canadian portioll of theextent of calving than when they used the Arctic Refugecoastal plain and the 1002 Area. Rcgardli:ss of timing ofsnowmelt in Canada, calving did there was dominated bymosses ;lnd evergreen shrubs (58.4-73.5%, Russell et al.1993). These forage groups were much less digcstibleIhan the iillmalllfe cottongrass llowers and willows(Russell et al. 1993) that dominated the calving diet of thePorcupine caribou herd in 1993 and 1994. This impliedIhal diet quality during calving was reduced when thePorcupine caribou herd IIsed the Canadian portion of theextent of calving rather than the Arctic Refuge coastalplain and the 1002 Arc;\.

Habitat Selection

Habitat sekctioTllllay be assessed at sever;ll orders(Johnson 1980); selection at each order impliesdisproportionatc usc of sOllie component(s) of the hahitatsthat arc available. For migratory barren-ground caribou,selection orders might be defined as follows frolll highestto lowest order:

First Order -thc specics distribution on earth.Second Order - area usc by herds within thc species

range.Thin! Order - annual range use within herd ranges.Fourth Order - se;lsonal range usc within annual ranges

of herds.Fifth Order - annual use within the aggregate extent of a

seasonal range.Sixth Order - annual concentrated u.->e within:UJ annu:Il

seasonal range.Severllh Ordcr - patch use within a conct.'ntratcd LISt.' arca.Eighth Order - pbnt specks usc within habitat patches.Ninth Order - plallt part usc withilfplant spccies.

Higher ordcr selectioll may nmstrain thl~ choices atlowcr orders (Johnson 19S0.J. Thc basi.~ of sclcctioll mayor may llot be eOllsisll'nt among orders and, when thebasis of selection changes among Mders, hahitat selcctionis considered to hc scak-dependcllt (O'Neil :JIld King1998). In this work, we assessed habiWl selection at fifthand sixth Mders as defined aboV<.'. l\luch discussion h;l.~

(oclI:',Ctl on fOllrth t}f(!L~r sclcclion (d. llergcrud alld Page

1987; Fr)'xell 1991, 1995), but analysis of selection at thefounh order for the Porcupine caribou herd was beyondthe scope of this repon.

For the purposes of the m;ltcrialthat follows, wedefine fifth order selection as the comparison of usewithin the annual calving grounds (ACG) to availabilityin tlle extent of calving (EC), written as ACGIEC(hereafler called cail,jllg grow/(f sdcclion). We definexi-t'th order sekctioll as the comparison of use withinanllunl concentrated c:llving arcas (CCA) to habitatavailability within the annual calving grounds (CCMACG, hercafter called cOIlCt'lJlraled cah,jllg seleclion).

Because Ihac was spatial dependency alllong habitats(vegetation, NOVI estimates, snoweover; all inventoriedfrom the same I-km' pixels) we present Ihe results foreach habitat attribute separately. Selection was assesst.'dby comparing mean usc/availability ratios among yearswith the null use/avail:lbility r<ltio of 1.0.

Habitat conditions within the extent of calving havebeen variable during 1985-2001. There was substantialsnowcover throughout the extent of calving in 1986,2000, and 2001, but greening was early inl990, 1994,1995, and 1998 (Fig. 3.18).

TIlere was scale dependency in habitat selection by thePorcupinc caribou herd during c<llving. Parturient femalessclected :lJlIIual calving grounds with proponionatelygreater area of high (>median) rate of greening(NOVLrate, 1.33x, p:::: 0.005) (Fig. 3.190) andproportionately less arca with high forage biomass both atcalving (NOVLcalving, 0.6Ox, I' < 0.001) (Fig. 3.19b)and during peak lactation (NDVL621, O.70x, p:::: 0.002)(Fig. 3.19c) than available in the extent of calving.

Panuricnt females also selected annual calvinggrounds with proponionatcly more area in the 26-50%(1.76x, 1':::: OliO!) and 51-75% (l.71x, p:::: O.008)SllOWt:over classes and proportionately less area in the 0­25t;~. (0.84x, P :::: o.onS) snowcovcr cbss than avail;lble inthe extent nf calving (fig. 3.20).

Analysis of vcgetation types in aJHltJal calving groundsshowed that parturient females selected wet sedge (I.42x,p", 0.004), herbaceous tussock tundra (1.42x, P < D.OOI),and riparian (1.37x, P < 0.000 \'cgctation types, avoidedthe alpine vegetatioll typc (0.60x, P < (lOOl), antI did notrespond (I' > 0.(5) to the shmb tussock tundra or moistsedge vcgetation typcs (Fig. 3.21).

In cOlltrast. at the next Iowa selection order (shth),p:lTturient fcmales of !lIe POTl::upine caribou herd selectedconcentrated calving areas with proportionately gn::alt::rarea of high roragt:: biomass bolh at calving(NDVI3alving, 2.35x, l' < 0.0(1) (Fig. 3.19b) and duringpeak lactation demand (NOVL621, 2.59x, P < 0.001)(Fig 3.19c) thall available in the annual (,alving grounds.Tht.' females were non-selective (I' > 0.05) for rate ofgret::uing (NDVI_r:lte) (Fig. 3.19a) and all SllllWcovcrcbsse,~ (Fig. 3.20), selecled herbaceous tllssock tllmlra

22 BIOLOGICAL SCIENCE REPORT USGSIBRD ::!(~J::!·f}{)()1

Cloud

Figure 3.18. Annual condilions of snowcover and vegelatiorl phenology derived from Advanced Very High Resolution Radiomeler (AVHRR)satellite imagery during the calving period (30 May· 5 June), 1985·2001, for the Porcupirle caribou herd. No concentrated calving was detected in2001.

! ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUto.1MARIES 23

'J+ ",- fl-o

'"£~"

J",j"",

(1.68x, P =: 0.001), avoided alpinc vcgetation (O.34x, P <0.001), and werc non-responsive (P > 0.18) to theremaining vcgctation types (Fig. 3.2]).

Although selection of vegctationlypes was scalc­independcllt, therc was scale depcndency in thc selectionof foragc quantity (NDVCcalving, NDVI_62I) andquality (NDVCmtc). Parturient Porcupinc caribou hcrdfemales selected annual calving grounds with a hjghproportion of easily digestible forage (NDVIJate), thenselected I:oncenlrated calving areas with relatively highplant biomass atl:alving (NDVI_calving) and on 21 June(NDVI_621).

Thc basis of habit,1l selection shifted from foragequality to forage quantity bctween the fifth (ACG/EC)and sixth (CCMACG) ordcrs. 'llle work of White et <II.(1975) and White ,lIld Tmdell (19801J) ,It the levels ofmicrohabitats (-seventh order, selection for biomass) andplant spccies withinlllicrohabitats (-eighth order,selection for digestibility) suggests that the basis ofselection continues to be dynamic ,lcross sliccessivelysmaller scaks.

For<lgc quality appears to be the ba;;is of selection atboth relmively large (fifth order) and reblively small(eighth order) scales. Forage quantity appears to be theb:l.~is of selection at interllledi:lh.~ scale.~ of analysis withinthi, range. Specification of the .'iC'11e of analysis is criticalto developing an understanding of the basis of f()rageselectioll by ungulatcs, and Porcupine herd cariboudemoll':;tratcd a variable functl(mal response In foragc(NDVl estim;Hcs) within thl' C.\!l'lll of calving.

Figure 3.20. Average percent of area in 4 exclusive snowcoverclasses for the aggregate extent of calving, annual calving grounds,and concentrated calving areas of the Porcupine caribou herd, 1985­2001. Statistically significant selection or avoidance (P < 0.05, overallexperiment) in comparison with the calegory to the left is indicated by'to or '." above the bars. For example. female caribou on the annualcalving ground avoided areas of 0-25% snowcover and selected areasof 26-50% and 51-75% snowcover when compared with availability inthe aggregate extent of calving. No significant selection of anysnowcover class was detected for the concentrated calving area whencompared with availability in the annual calving ground.

+

+

r~DVl.-e21 Ci...

,,",,",I c.r,;n~ Grou.",d • ee,-.<Mt..,lO<l Ca',;".

+

+

bJM

"M,

5:.0,&10;!~,""0

oj

Figura 3.19. Average percenl of area in low « median} or high (>

median) classes of a) daily rale of increase in lhe NomlalizedDiffereflC€ Vegetation Index (NDVUate) b) NOV) at calving{NDVCca!ving}, and c} NOVI on 21 June (NDVI_621) for the aggregateextent of calving, annual calving grounds. and concentrated calVingareas of lhe Porcupine caribou herd, Alaska. 1985·2001. Statisticallys"lgnificanl selection or avoidance (P < 0.05, overall experiment) incomparison with the category to the lell is indicated by 'to or'-' abovethe bars. For example, female caribou on the annual calving groundavoided low NDVUate and selected high NDVUate in comparison\'.ilh availability in the aggregate extent of calVing. No significantselection of NDVUate for the concentrated calving area whencompared w;,th the annual calving ground was detected.

lE----" h1llh-r~DVI_c'M~~ CI•••

[J&':~~l01 Cl>r,i"1l ElliJ Mou.1 C'M"~ Grwnd • eo.-.c"nu.:O<l C'~'\">J

24 BIOLOGICAL SCIENCE REPORT USGSfBRD 2002-0001

•

Figure 3,21. Average percent of area in 6 vetetation types for theaggregate extent of calving, annual calying grounds, and concentratedcalving areas of the Porcupine caribou herd, 1985-2001. Vegetationtypes: Wsedge::: wet sedge; Msedge::: moist sedge; HerbTT:::herbaceous tussock tundra; ShrubTT::: shrub tussock tundra. Alpine.and Riparian. Statistically significant selection or avoidance (P < 0.05,overall experiment) in comparison ....'ith the category to the left isindicated by "+. or·o • above the bars. For example. the female caribouon the annual calving ground avoided the Alpine vegetation type andselected the HeroTT vegetation type when compared with ayailabilityin the aggregate extent of calving. and on the concentrated calYingarea the caribou showed similar selection when compared v.1lhavailability in the annual calving ground.

lllcrc were no clear difference~ in pallerns ofselection of any types of h:lbitat~ between the incre:lseand decrease phases of the herd. 11lis observation istempered by tlle fact that habitat selection was assessedfor only the l:lst5 years (1985-1989) of the iJlcrc:l.~e

phase, but has bC!:1l nssessed for nlll2 years oftllecurrent decline (1990-2001).

111e shifting location of annual calving grounds withinthe ex lent uf calving was apparently a functionnl respouseto :lllllually variable landscape pallerns in the qualllity ofeasily digestible forage (NDVCrate). The location ofcOllcentrated calving areas within annual calving groundswas an appnrent function;11 rcsJlonse !O forage biomass(NDVLcalving, NDVI_621).

lllis functional response to h;lbitals allowedPorcupine. caribou herd fcmales to allain suhstantialintakes of nitrogen (Fig. 3.22) based 011 estilll;\ted dietcomposition (Figs. 3.16(1, 3.170), estimated nitrogencontent of consumed forages, and consumption ratespresemel! by White ct ;11 (1975), White and Truddl(19801/./1), and Trudell and While (1981). -lllUS, IhePorcupine caribou herd calving ground was clcarlyimportant to the ;\nnllal nitrogL~1l budge! of I;lctalingrcmales and was likely important 10 the annual energybudge!.

The :ldjacent Central Arctic herd obtained only aboutolle-quarter as llluch dict;lry nilrogen from its calving

o lcc'-_-c:-L_; l L '-_L__~-'-CIlnlrnl Arctic POlcupino Bolhursl Goorgo Rivor

Caribou Herd

Figure 3.22. Estimated 10tal intake of dielary niltogen (9) from !hecalving ground (25 May - 14 June) for 4North American caribou herds.Forage composition of diet and nutritional composilion of forages wereestimated from locally collected samples. Intake tales were estimatedItom White el al. (1975).

ground as did the Porcupine caribou herd (Fig. 3.22). It islikely that the proportion of the annual nitrogen budgetobtained from a cah'in£ ground is positively correlatedwitll the relative value of the calving ground to tIlenutrition of a herd wilhin its annual range.

Effects of Insect Harassment on Habitat Use

1\-10s(]uitoes (Cllcu/ida~') and flies of thl': familyOestridae arc known to harass carihou. althoughh;uasslIlent by Oestrid flies Illny OCCllr primarily afterPorcupine herd caribou leave the calving ground.Lactating females that are disturbed by insects mayexperience a negativc energy balance due to increascdmovelllent rates when trying 10 escape harassment byinsects (White et :d. 1975. RlISseli et al. 1993). Whenharassment causes lactating felllnlcs to substantia.llyreduce foraging time, calf growth lIlay be reduced (Helleand Tarvainen 1984, Fancy and White 1987, Russell et al.1993).

During wann and calm days (mean temperature> IJoeand mean wind speed <6m/sec) when conditions weresllch that caribou were likely harassl'd by illsect~ (Nixon19(0), Porcupine herd caril)()u prefcm~d dry prostrateshrub vegclation types on ridgc tops in the fOOlhills and1ll0lHIlains of the Brooks Range, clcv:lled siles on thecoastal plain, and areas adj:lcent 10 the Beaufort SeaC(Ja.~t, apparently lO gain relicf frolll f]losqui(()es (Wabh elal. 1992).

Porcupine Ill'rd caribou did !lot display as strong atendency to move to the coasl!ine during plltclltial insecthnrassillent as has been seen for tile adj;K:clll CentralArctic herd. Observation.~ of Ill\JVCl1lellts of unmarkednnimals during survey flights, howcver, indicate that

,t\Rcnc REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU~Il\IARIES 25

segments of the herd oflen follow the coastline whilemoving along the coastal plain of the Arctic Refuge inJuly (F. J. Mauer. U.S. Fish and Wildlife Service,person::li cOllllllllniculion).

Individual radio-collared caribou showed at leaslpartial fidelity (Le., caribou repeatedly returned tospecific areas) to either the coastal plain, foothills, ormountain zones during the insect harassment season indifferellt years (Walsh ct al. 1992). 111C negative energeticconsequences of insect harassment (I-Idle and Tarv:lincn1984) suggest that frce access to insect rdicfhabitat isimp0l1:1111 to caribou (Walsh ct al. 1992), but in sOllieherds the energetic cost of insect harassment may be low(Toupin et :II. 19(6).

11 ~,

i I' II ~I~, liil !<

I~f~ I' 1"'1';11'III' IIII'", I' I ""'II II'H[ 1,111

,'111'11 '\",::

1 1',,1,'1 J fill, 'Iili d ,- 1993' ---- 1994

Year

[JQ-.3wccks

rmniWl4-5 weeks

iII

J:1

j·1j

I

Calf Performance In Relation to Habitat Use

t.lean calf weights wilhin 1-2 days of binh \liereremarkably similar among years. On averagc, femalecalves caught during 1992-94 when the herd wasdeclining weighed 6.2 kg, slightly less (I' = 0.003) than~;;2-day-old female calves caught during 1983-85 (6.7 kg,Whitten et al. 1992) when the herd was increasing.

The increase/decrease classification, however,explained only about 9% of the variance in calf weights.111e difference in female calf weights between theincrease and decrease phases of the herd was due solelyto a cohon of heavy calves in 1985 (7.2 kg). Femalecalves caught in 1983-84 weighed an avcnlge of 6.3 kg(Whinen et al. 1992).

There was a significant interaction among years andbetween Jleriods (O-3 weeks and 4-5 weeks after birlh) (I'< 0.001) in daily weight-gain of fcm:de c;llves, 1992-94(Fig. 3.23). Daily gain wa's panicularly low during tIlefourlh and t1flh weeks of life for calves born in 1993 (Fig.3.23).

Daily weight-gain of CDlves did not differ betweencalves born in the concentrated calving are~IS and inlheperipheral calving areas (I' = 0.214). ~Iuch higher relativedensities of caribou (7x. on average) in Ihe concentr:l!edcalving area~ compared to peripheral calving areas lll;lYhave reduced forage available to individuallacta!illgfemales.

Even tllOUgh cOllcelllrated calving areas fwd a greaterproponion of area with high plant biomass (bothNDVI_calving ;llld NDVC621) than did the annualcalving grounds, the differential ill forage abundance wascvidellily not sufficiellt to overcome the higher densitiesof caribou in the cOJlcentrated calving :In,'as and toenhancc the weight-gain of cal ves horn there.

Paltcl"lls of habitat use by calve.~ var-icc! signit1c:llltlyif> < 0.(1) betwcen periods :llld ;\IlHlng year.~', 1992-199-1(Fig. 3.24a·(;), but were generally similar to use of sitesfor calving (Fig. 3.21). Weight-gain of cal\'cs duringcalving ground usc was 11ll! :lssociated with the percell! of

FIgure 3.23. Daily gain (kg) of caribou calves of the PorCtJpine herd,1992-1994, durin9 2 periods (O-3 weeks post·birth and 4·5 weeks post·birth). Gain was estimated from sequential weights or recaptured radio­collared anImals. Means are listed above the appropriate bars.

time that calves spent in any particular ve,getalion Iype orin any class of forage at calving (NDVCcalving), rate ofincrease in forage during lactation (NDVIJate). fOnlgeavailable at the peak of lactation (NDVC621), or~nOwcover(P > 0.05).

Although individual calf weight-gain was notexplained by within-annual-calving-ground habitat usc,several characteristics of parluricnt females and calveswere rclated to habitat conditiolls in the anllual calvinggrounds, 1992-1994. 111e rank orders of I) NDVC62I inth~ annual calving gwund, 2) average parturient fcmaleweights (Fig. 3.25), 3) p:lr1urient femalc'hody conditionscorc, and 4) average calf weights, all at 3-weeks post­calving, were all the sallle (1993 > 1994> 1992).

Lack of correlation betwccn indi vidual calf weight­gain and usc of annual calving ground habitat suggeststhat the location of annual calving grounds Illay havemaximizcd calf weight-gain, given the conditions of theannual habitat avuilable within the ex.tent of calving. Oncethe annual calving ground WHS located in an area thatprovided a high propUrlion of easily digestible forage(high NDVI_fate), then variation in caribou density andforage biomass (NDVI_calving, NDVL62I ) may haveinteracted to reduce vari;l\ion in performance among theindividual study animals.

Factors Associated with Calf Survival on theCalving Ground

During 1983-19H5, average 1ll0l1ality \)f calves duringJune was 29';,,;; (Whitten et al. 1992), slightly higher thanthe 1983-2001 avcrage of 25';';". III those early years, about61 % of mortality on the calving ground was due 10predation antlt!ll'. !L'll1ainder 09%) was due to llI11ritional

26 BIOLOGICAL SCIENCE REPORT USGSIDRD 2002-0001

or physical charac{~ristics of calv~s (Whitten et al. 1992,Roffe 1993). 'nle interaction betwe~Jl nutritional status ofthe calv~s and pr~dation mortality was not known.

Figure 3.24. Availability of 6 vegetatiol1types in the aggregate extentof calving for the Porcupine C<lrioou herd and use by radio-collaredcalves durirlg 2 periods (0-3 weeks post·birth and 4·5 weeks post-birth)for (1) 1992, b} 1993, and c) 1994. Vegetation types: Wsedge =wetsedge; ~,lsedge:: moist sedge: HerbTI =herbaceous tussock tundra;ShrubTT =shrub tussock tundra, Alpine, and Riparian.

'" ] r:,030 - 1-, I- E0 028 "

f 0213 ,1 i "eo 5,B -,

0.24 - -is N9 ...

I"022 - " ;;<020 •;:, " ~

~, 0.18 I ~

2 [) 113 ~n

, ;0\4; 1.._--- "1S92 199J 1994

Y~3'

CI r·:DV'_"~1 •FCm.l!<"M . ~1 JlJ""

Predation occurred further south and at higher elevationsncar the foothills during 1983-1985 (Whilten et al. 1992).

During 1983-1985, golden eagles caused mostpredation mortality of calv~s on the annual calvinggrounds (-60%), grizzly bears mnked second (-24%),and wolves ranked third (-16%) (Whillell et al. 1992).Young and l\1cCabe (1997) estimated that bears killedabout 2% of calves during 1994, a year with rclativ~ly

high overall calf survival (Fig. 3.1 Ob).Immature golden eagles ranged throughout the coastal

plain and foothills (Clough et al. 1987), while goldencagle nests and wolf dellS were primarily restricted to thefoothills (S£'I' Fig. 6.1). Grizzly bear densities weremoderate and their distributions were concentrated in thefoothills (Voung and McCabe 1997). In late summerthrough winter, the source and distribution of predationmortality of calves were unknown, but wolves wereprobably the dominant predator.

We Ilsed mUltiple scales to analyze factors associatedwith calf survival during June: I) fate of individual calveswithin the population of calv~s; and 2) the proportion ofthe annual popUlation of calves that survived Illltilthe endof June in relation to a) habitat characteristics within theextent of calving and b) habitat characteristics wiUlin eachannual calving ground. These latter 2 classifications areconceptually equivalent to the fifth and sixth order habitatselection analyses.

Several factors were associated with enhallc~d

survival of individual calves, 1983-1994 (n = 345 calves).Survival was greater (10.8%, P = 0.004) if the calf wasborn in a high density concentrated calving area ratherthan in the low density peripheral portion of the calvingground; greater (11.0<:"0, I' = 0.008) if born ncar themedian calving date rather than being born early or late inthe calving season; greater (I 1.2%, P = 0.006) ifbom on

Figure 3.25. Median Normalized Difference Vegetation Index on 21June (NDVI~621) withirllhe annual calving grounds of the Porcupinecaribou herd arld weights of parturierlt female caribou when C<lplured,....ithin the anrlual calving ground Orl 21 JUrle, 1992-1994.

I"~

i".'

"I;rjiiII _J1._"L,llHort>n s ....ubn ",'p;.,,, R'~,,"~

Vo>g<)!aiJoo Typ<>

DM<l'I"to/<l fS']USC.WC'CksO.J • lJ'".o·WC'Cks 4·5

'WJ

VClletatlOll T,pc

[]/'V~,i"~IC !i0]L,':;C.W""ksO.3 .USC.WC'<lks4.:'

""~~J(I

"

till co.. r+.!';llt,~I'_LGtI'o'>o<lo;)o "'_00 H.r~ 5rv'ubIT Alp''''' R,P'l"""

0)

bJ

oj

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 27

,.,O~ O~ OM o~

NOVI_621 . Extent of c~tYing

Figure 3.26. Call survival through June for the Porcupine caribouherd, 1985-2001, ill relation to median Normalized DifferenceVegetation Index on 21 June (NDVC621) .....ithin the aggregate exlentof calvillg (EG). Legends identify the year of the estimate. Calf survivalwas not estimated in 1986 because inclement weather prevented acomplete sample in fate June. Calf survival for 1993 was a significantouUier (RStudellt = 3.84, see lext for biological justification) and wasexcluded from the estimated regression lill8 {f' =0.85, P < 0.0001).Upper and lo.....er dashed lines indicate 95% confidence intervals on thepredicted observations.

modd can be used to assess whether calf sun'ival duringJune is affected by development.

Calf sun'ival for 1993 was an outlier (RStudent =3.84) and excluded from the estimated relationshipbetween NDVI_621 in the extent of calving and calfsun'ival (Fig. 3.26) alld from all subsequent models ofcalf sun'ival. During 1992, :ltlllospheric aerosol.~ from theemption of fl.lt. Pinatubo in the Philippines re'1clled theArctic ill the spring (Stolle et a!. 1993). 111is resulted in alate spring, cool SUIlUllt:r, early and heavy snowdeposition in the fall, and ncar catastrophic conditions forcaribou.

We sunnise tllat the consistently bad weathcrconditions during 1992 and early 1993 resulted ill a carry­over effect t1Klt reduced calf sun'ival in 1993 to levelsmuch lower than would havc been expectcd on the basisof NDVl_621 alonc. It was likely that this suspectedadditional mortality in 1993 affected calves within thefirst day or two of life; perhaps many calves were of verylow binh wcight. We draw this conclUSion because 0- to3-week weight-gain of calves that sun'lred to be radio~

collared ill 1993 was as high as any other year (Fig. 3.23)and the weights of parturicnt females Ihat were caughtwith their live calvcs on -21 Junc in 1993 were as high asany weights we observed, 1992-1994 (Fig. 3.25).

At the smaller scale of the annual calving grounds, theproportion of Porcupine caribou herd c:llves that survi\'edthrough June was positively related to both NDVC621 inthe annual calving grounds and to the proJlortion of calvesthat were born all the coastal plain (assumed lower

Ihe coastal plain with lower suspected density of wolves,eagles and bears; and greater (8.3%, P = 0.026) if born inthe 1002 Area.

The survival advaillage of high density calving toindividual calves tended to be greater when calves wereborn in tIle foothills and mountains than when lhey wereborn on the coastal pl:lin (14.3% advantage vs. 7.9%advantage, respectively).

Individual calf survival was not related (P = 0.160) tothe fre(jueney of use of its binh site as :l ponion of theeonecntrated calving area, 1983-1994, but calf survivalwas 100~:er (9.9%, P = 0.026) if the binh site was in anarea n4'er used as a concentrated calving area. In astepwise logistic regressioll analysis thai simultaneouslyconsidered calving density, tillle of binh, zone of birth(coastal pl:lin or foothills), and in or out of the 1002 Area,only calving density (P = 0.004), time period of binh(early, middle, late; P = 0.(12), and zone (1' = 0.(08)entered Ihe lllodellh:ll predicted individual calf survival,1983-1994.

'111C surviv:lI advalllage of both high calving densityand being born near the middle of the calving period mayhavc been due to predator swamping where high spminland temporal densities of calves may make it difficult forpredators to capture individual calves (Hamilton 1971).Bears tended to be less successful at capturing calves inthe concelllrated calving areas of the Porcupine caribouherd (Young :lnd McCabe 1997).

When assessing the proponioll of the unnualpopulation of calves that surVived during June, the limingof binh in relation to other calvcs W:1S not applicable, butmedian calving date, J983-1996, was available. Inaddition, we could consider Ihe relativc alllount of food(NDVCealving, NDVCrate, and NDVL621), winterrange conditions prior to calf binh (snow propenies), andthe proportion of calves horn in coastal plain or foothill,",ones.

Analyses of the proportion of calves surviving inrelation to these independent variables were conductedscparately :"It 2 scales: a) the extent of calving and b) theannual calving grounds.

Within the extent of calving, the relative amollnt offorage available to females during peak lactation(NDVL621) provided the best mouel of calf sun'ivalduring June (r = 0.85, l' < 0.0(1) (Fig. 3.26). No otllerindependellt variable that was considered addedsignificanl explanatory power.

This model (Fig 3.26) (Percent JUlie Calf Survival =10.107 + (2.U:) • NDVUi21 in the extent of calving)] ~

JOO) was the best available cstimate of survi\'al of calvesduring JUllC for the Porcupine caribou herd underundisturbed conditions during the past 2 decades. '111ismodel of calf sun'ival was independent of annual calvingground location and, if the 1002 Area is developed, the

28 BIOLOGICAL SCIENCE REPORT USGS/ORD 2002-0001

predalion risk) (r = 0.70, P < 0.00l). No other vari;lbleadded significant explanawf)' power. Median NDVI_62Iin lile annual calving grounds and the propof1ion of calvesborn on lhe coastal plain were nOl correlated (P> 0.94).Forage in the 3Jlnual calving ground accQullled forapproximately 75% of the total variance explained by thismodel and assumed pred:llion risk accoullled for theremainder (Fig. 3.27).

TI1US, in addition to scale dependcncy in the functionalresponse M caribou lo habilats (sc!eclion of NOVIswithin the extent of calving and within the annual calvinggrounds), there was scale dependency in the numericalreSJlonse of calf survival to calving ground location andhabitat conditions. Only forage \\,;15 related to calfsurviv;ll at the l;lrgesl spatial scale (extent of calving) thalIVe :lnalyzed.

At the intermediate scale (annual calving ground),Corage dominated calf survival, bul predmion risk addediubstantial explanatory power. Attlle snlal!est scaleindividuals within the population of calves), spatial andemporal v:lriance in calf density (indirect predation risk)md direct pred:l!iOll risk 1I10st effectively explained calfurvivaJ.

TIlis scale dependency ill calf survivallikeJy occurred'ecause the annual vari:lllce in habit:ll conditions in bothhe eXlcnt of calving and in the annual calving grounds farxceeded the annual variance in predation risk within thextcnt of calving and \vithin thc annual calving grounds.'he scale dependency in caIr survival made it impossible) extr..lpolate across scales. Thus, 10 develop annder$landing of the relativc influence of forage and

ure 3.27. Predicted calf survival for the Porcupine caribou herd,\5-2001, in relation to median Normalized Difference vegetationex on 21 June (NDVU21) within lhe annual calVing ground and 10proportion of calves bom on the Arctic National Wlldl'lfe RefugesIal plain physiographic lone where predator densily \'las lower1in the foothill-mountain physiographic zone (r"::: 0.696, P <)1). Calf survival was nol estimated in 1986 because inclement,lher prevented a complele sample in lale June.

predation on calf survival, it is imperative 10 specify thescale of analysis, and nssess multiple scatessimultaneously.

'nle temporal increase in forage during peak laetalion(NOV,-62I ) (Fig. 3.4) was coincident with local climatewarming (Fig. 3.3a). Forage at calving (NDVCcalving)was positively associated with the Arctic Oscillation (Fig.3.6). TIlere were also pOsitive relationships betweenclimate and NDVl3alving, between Jlercent of femalescalving in the 1002 Arca and NOVI_calving, and betweencalf survival alld NOV I_calving [t..l = 0.33, P = 0.011(annual calving ground); r = 0.60, P < 0.001 (extent ofcalving)]. As a result, June calf sllrvival was weaklycorrelated (r = 0.22, P =()'029) with the proponion ofcows that calved in the 1002 Area. Further, becauseclimate affected calving ground location (e.g., Porcupinecaribou herd females were more likely to use the westernpOf1ion of the extem of caldng following winlcrs with apositive Arctic Oscillation), both forage availability andpredalioll risk were implicitly related to climate.

In YC;ITS with substantial Sllowcovcr on the coastalplain (Fig. 3.18) and rekuively low NDVI_62I in theextent of calving, average calf survival (66%, /l =7, SE =690) was 19% less (P = 0.0(8) than when there was littleS!lOWCOVer at c:Jlving and NDVC621 was high (85%, fl =6, SE = II %). Thus, climate was an important influenceon habitat conditions, on tile likely lise of the Alaskacoastal plain and 1002 Are;l for c;llving, and on calfsurvival during June, 1983-200 I, ullder undisturbedconditions.

Potential Effects of Development on June CalfSurvival

In ordl'r to assess the polelllial cffecl.~ of devdopllletllof the 1002 Arca on thc Porcupine caribou herd duringcal ring, we needed a model of caribou behavioralresponse to oil field infrastruclilres. 'I1le adjacent CentralArclic herd (Fig. 3.2), which calved in lhe vh;inity ofPrudhoe Bay - Kupal1lk complex of petroleumdevelopment arcas. provided lhc only available model ofcaribou behavioral response to petroleum de\'dopmelllduring calving.

PIlr{/lrielll ji.'IIIll!e caribou (i.c., those about to givcbinh or accompanied by very young calvcs) of the Celllral.·\rclic herd repeatedly Jemostraled their scnsitivity toJislllrb;lIlce during the first few wed;s of life of lheircain's (Smith arld Cameron 1983, Whittcn and Cameron1983, /)au and Cameron 19~6: Cameron et al. 191)2;Ndlemann and Came['{)11 19%. 1998).

Parturient fcmales ;l\"oided, \)l" wcre k.~s likely tocross. inji'ilSlntt'llires (roads and pipelines) during thecalving SeaSl)lj (Cameron and Whillen 1979, D:lIl andCameron 1986. !\IUfphy and Curatolo 19H7, Ll\vhcad1988, Cameron et ;iI. 1(92). In ;ldditioll, densities of

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 29

III

caribou during calving (June) were gre<lter than expectedbeyond 4 km from roads and pipelines (Cameron et al.1992).

Central Arctic herd caribolllllay make substantial uscof areas in the vicinity of oil field infrastructures duringperiods of moderate to high insect abundance during P{)st~

calving in July (Pollard et al. 1994). 11lat observation isnot relevallt, however, to the distribution of thc CentralArctic herd during c:llving ill JUlle nor to the assessmcntof Porcupine caribou herd distriblllion during calving illrelation to potential oil development: Caribou of thePorcupine herd generally depart the calving groundduring early July.

Historically, 2 zoncs of concentrated calving of theCentral Arctic herd havc becll recognized (Murphy andLawhead 2000). The zones were physically divided by thcSagavanirktok River and the trans-Alaska oil pipeline.nlcre was an e:lstern reference zone where developmentinfrastructure was historically :Ibsen! through 1995, and awestern deVeloped zone that included the Prudhoe Bay,~lilne Point, and Kuparuk petroleulll developmem areas.In 1996, the developed versus reference zone studydesign was compromised by the completion of pipeline!;le:ldin£ to the Badami petroleum development area, castof the tran!;+Alaska oil pipeline and into the referencezone.

During the late 1980s, concentrated calving in thedeveloped zone shifted from the vicinity of the Kuparuk­l\·filne Point petroleum development areas to undevelopedare<l~ to the !;outh-southwest of the oil fidds (Lawhead etal. 1993, Murphy and Lawhead 2000). Low densityc:llving continued 10 occur in these petroleumdevelopment areas while concentrated calving shifted.'Illat shift was completed by appro:\il\l:llcly 1987 whenthe Oliktok Point and r....1illle Point roads were completedand substantial infr:lstructure was in place. nle uni­directional shift in concelltr'Jted c:llving in the deVelopedzone, 1980-1995, has sUbSe{jllently been confirmed (P <0.002, Wolfe 2000). During the same years, however, theconcentrated calving area in the reference area showed noun i-directional shift (P = 0.14, Wolfe 2000) (st't' also Fig.4.7).

Since 1996 the bulk of high densily calving in thedeveloped zone has remained south of roads and pipelinesalthough a slIlall zone of high density calving occurred intlle Kuparuk-Milne Point :m::a in 1996 (Lawhcad andPrichard 2001). The shift in calving distribution in thedeveloped zone occurred even though the Milne Point andKllpamk pctrolcllJn development area~ included:;Ub.~l:llllini illlpro\"emellls in field design and layout (e.g..elevated pipes. reduceu road l!cnsity) tli;l! should havefncititated caribou passage clllllpared with the dc~igll ofthe older Prudhoe Bay COlllpleX.

No olher concentrated calving ;lrca of Alaska harrcll­ground herds has demonstrated a ~latisticaJJy significant

uni-directional shift during the past 2 decndes.Kelleyhouse (2001) showed no un i-directional shift inconcentrnted calving for the Western Arctic herd, 1987­2000, but was unable 10 asses.~ shifts in the concentratedcalving areas of the Teshekpuk L:lke herd due to aninadequ:lte number of ycars forthe test. As notedpreviously, directional shift~ of concentrated calving areasof the Porcupine caribou herd have not differed fromrandomness, 1983-200 I.

Forage during peak lactation (NDVC621) in thccOllcentrMed calving area in the developed zone of theCentral Arctic hcrd declined as the concentrated calvingarea shifted south-southwest, 1980-1995 (Wolfe 20(0).During this shift, forage during peak lactation remainedhighest in the area llsed for concentrated calving during1980-1982 (Wolfe 2000). nlere was, however, no declinein forage availability on June 21 (NDVI_621) in theconcentrated calving areas in the reference zone of theCenlral Arctic herd during 1980-1995 (Wolfe 2000). Noclear biological evidence explained the shift ofconcentrated calving in the developed zone to an area ofreduced forage availability for lactating females. 11lUs,petroleulll development was implicnted as a cause of thesoutherly shift in cOllcentrat.::d calving in the developedzone of the Central Arctic herd, 1980-1995.

Since the first census of the Centml Arctic herd in1978, the herd size has increased from approximately5,000 to :lpproximately 27,000 animals in 2000 (E. A.Lenart, Alaska Department of Fish and Game, personalcomlllunication. Sa also F(r; 4.2). 'nlere was a sharpdecline (from 23,000 to 18,000) in the herd from 1992­1995 and a subsequent recovery. It is unknown whetherthe Centr;ll Arctic herd would have increased at a higherrate than obsen'ed had the concentrated calving area inthe dcveloped zone not shifted 10 the sOllth-sollthwest by1987.

'111e ohsen'ation of either;l/l inCl"eas.:: or decrease ofany magnitude in the size of the Central Arctic herd orany othcr herd is not, by itself, sufficient evidence toconclude that there has be.::n an effecl of development orlack thereof on herd size. For example, had the 1002 Areabeen developed in 1989, the sub~eqllent natural decline ofthe Porcupine caribou herd (Fig. 3.8) would not haveeonstitutcd evidence of an dicci of devdopillellt.

To asscsS potential effects of devt::!opillent on thegrowth curve of the Ccntral Arclic herd, we needed t{)make cOlllpari.~ons with an ecologically similar Ill~rd. ThePorcupine caribou herd docs not con~titlile a goodl.'cological cOlllparison and neither dlws the WesternArctic herd. "nle Teshekpuk Lake herd (Fig. 3.9) is themost ecologically comparable herd to thc Central Arcticherd in Alaska.

Thl~ C~ntral Arctic herd and Tesllekpuk Lake herd arecertainly not identical, however: I) bolh herds arcrelali vely small ill size and the trajectories of their growth

30 BIOLOGICAL SCIENCE REPORT USGS/BRD 2002-0001

curves suggest exponential growth, 2) both herds haverelatively high bull:cow ratios (-80: 100),3) calvingground habitats of both herds showed similar climatetrends (Kelleyhouse 2001, Wolfe 2000), 4) both herdsexhibited the same dip in herd size during the rnid-1990s(Fig. 3.9), 5) neither herd has consistently demonstratedlhe long distance migrations exhibited by the Westernr\rcric herd and Porcupine caribou herd, and 6) before1987, both components of the Central Arctic herd as welllS the Teshekpuk Lake herd calved in wet coastal habitatsNith relatively late snowmelt.

The apparent divergence in the relative sizes of ~Ie

::entral Arctic herd and adjacent lcshekpuk Lake henllrter 1987 (Fig. 3.9) suggests that the growth r;He of the::elllral Arctic herd may have slowed after roads andlipeIincs expanded in the developed zone and theoncentrated calving area in the developed zone shiftedouth-southwest. The relative trajectories of the 2 herds'Towth curves were par<lllel through the mid- to late­980s when both herds were slightly less than 4 times asuge as when first ccnsused. 111en:=after, their trajectoriesivergcd slightly. By the late 1990s the 'Icshekpuk Lakeerd W;lS about 7 times larger than when firs! censused'hile the Central Arctic herd was only about 5.4 times aslrge as when first observcd. Cronin ct al. (1998) notedlat exponential growth rate of the Teshekpuk Lake herdas approximately twice as great as the exponential•:owth rate estimated for the Central Arctic herd (0.152J. 0.077, respectively) from the mid· I 970s through theid- I 990s.

Sevcral ecological factors llIay have diluted orlscured any population consequcnces of avoidance of'Hokum development areas by the Central Arctic herdIring calving. FiN, only the half of the herd that usede developed lone was Jlott~l1tially affected. Reduction in"aibble food for lactating females during peak lactationlS demonstrated only for the females that used theveloped zone concentrafcd calving area (approximately% of all felllales ill the Central Arctic herd; Wolfe00).Second. the Central Arctic herd remained on the

:lstal plain when it shiftcd its concentmted calving areasthe developed zone. The parturient females and calvesre not displaced !O the adjacent foothills where:dator densities were assumed to be greatest. 111US, thef! lllay have incuITed little if allY additional Jllortality:= to prcdation.'lllinl, developlllent of the complex of petroleulilIclopmem areas frolll Pl1ldhoe Bay to Kuparuk hilS:urrcd during a period of relatively favorable'ironmental cOlldition.~ U.,.faxwcIl 1996). Thc re.~ilicllcc

lerds I\l abiotic, biotic, or anthropogenic dl<lllenges'.lltl be expected to be gre;llest during favorableironlllental conditions.

Fourth, because thl: Central Arctic herd obtained arelatively sma]] proportion of its annual nitrogen budgetfrom its calving ground compared with other herds (Fig.3.22), the Celllral Arctic herd calving ground may havehad less relativc value to herd performance than thecalving grounds of other herds.

Fifth, calving ground density of the Central Arcticherd has been, and remains, quite low (npproximatelyone-fifth thl: effective density of the Porcupine caribouherd; Whitten and Cameron 1985). 'nms, even thoughfemales of the Central Arctic herd in the developed zoneshifted their concentrated calving to an area with reducedtotal forage, the amount remaining per caribou may haveheen sufficiellllO accotnmodmc nutritional requirements.

Because ecological conditions for the Porcupinecaribou herd are substantially different than for theCentral Arctic herd, il is unlikely that all theseameliorating factors will apply to the response of thePorcupine caribou herd to development within its calvingground. Ncvertheless, the avoidance of oil field roadsand pipelines by parturient females of the Central Arcticherd during the calving SC;lson is transferable toPorcupine caribou herd because sensitivity to disturballceby parturient caribou has been repeatedly noted clsewhcn:=(Wolfe et al. 2000).

To assess tlle potential effects of petroleumdevelopmellt in the 1002 Area on the Porcupine caribouherd, we assumed that displacement of Porcupine caribouherd's concentrated calving grounds would occur, similarto the shift observed for llle concelllrated calving, area inthe developed zone orthe Central Arctic herd (Lawheadet al. 1993, Wolfe 2000). We then used empirical habitat­demography relationships developed in the Porcupinecaribou herd studies to assess the implicntions of thishypothetical displacement on calf survi va] during June forthe Porcupine caribou herd.

We based our predictions on an empirical Illodelrelating calf sllfvivalto forage in the annual cah'ingground on 21 June and to the proportion of calves born inlow predation risk (Fig. 3.27). This empirical Jl\odel wasPercent June Calf Survival::: [-0.0396 + (2.0989" medianNDVL621 in the annual c;lh'ing ground) + (0.00283 ..proportion of calves hom in low pred:ltion risk)] .. 100,(,-2::: 0.70; P < 0.001). 111e spalially explicit nature of thisintermediate-scale lllodcl subsumed the effects oftemporal <ind spatial caribou density on individual calfsurviv:ll.

First, we IJsed the empiricalmoJd 10 predict calfsurvival in each of the 17 ()b.~ef\'ed allnual calvinggrounds oflhe Porcupinc caribou had, 19l'i5-2001 (Fig.3.13). Thcn each cOllcentrated calving area wa.~ displaccdthe minimum distance necc.~sary to pf{)vide 4 kill

clearance froill the boundary l)f e"Jch of 4 hypothetical oildevdopment scenarios for thc 1002 Area presented inTussing and I-laky (1999; scenarios 2-5) :llld for the

- ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 31

single hypothetical development scenario presellted in the1987 Finnl Legislative Environmental Impact Statement(Clough cl aJ. 1987). The scenarios in Tussing and Haley(1999) <lrc based on the most recent estimates of thedistribution and quantity of oil reserves within the 1002Area (U.S. Geological Survey 2001).

This protocol assumed oil field design similar to theKuparuk and /l.Elne PoilU petroleum development areaswithin the scenario boundnric$. The Illodeling exercisecoutd be llsed 10 assess the potential dfccl:> of additionaldevelopment scenarios that arc llot presented in Tussingand Haley (1999) or Clough C[ 011. (1987).

CCllIral Arctic herd parturient females actuallyseparated their concentrated calving areas fromdevelopment infrastructure by about 7-8 km (Wolfe2000). We used a conservative displacement of 4 kl1lbased on observations by Cameron et al. (1992) ofincreased caribou density from 4 kill outward beyondroads and pipelines. Calving sites and the entire annualc31ving grounds were displaced along with theconccntr3ted clliving areas.

Our protocol stated that a cOllcentr,l!ed calving areacould not be moved onto the Beaufort Sea. \Ve made nochrmges in shape of the concentrated calving areas orannual calving grounds. As a result of these shifts,relatively small portions of the peripheral, low-densitycalving areas were occasionally moved onto the BeaufortSea along with some associated calving sites. We treatedthese ocean sites as missing data when assessing thepotential effects of displ,lCcment Oil calf survh'al.

!'v1odcled displacelllclll for the Porcupinc caribou herdwas to the cast and sOll1h, parallel 10 thc Beaufort Scacoastline, because that is the direction of the herd'smigratory appronch to the ;lnnunl cnlving grounds inspring. Displal.:emellt ()f the developed-zone concentratedcalving nreas of the Central Arctic herd has beenprimaril y to the south, the direction of ~lpproach to thaIcalving ground from winter range.

Our protocol minimizcd di.~placcrnent of thePorcupine caribou herd calving grounds into the foothillsand mountain mne. This tcnded to kcl.:p thc ;mnualcalving grounds on the coastal plain in the best remainingforaging habitats. In sOllie cases, observed cOllcentr<t!edcalving areas (e.g" in 1988,2000, and 2001) did 1I0toverlap the boundaries of any of the hypotheticaldevelopmellt scenarios, and in those cases the annualcalving ground was not displaced.

Once the concentrated calving areaS and associ:l\edanllual calving grounds and calving sites \WrL~ displaced,the fnrage during peak lactation (NDVC(21) within thedisplaced annual calving ground was rt"iuventoried, themcdian was rl.:calculated. and the proportion llf calvesborn in the low predation risk zone (coastal plain) wasrecalculated.

.~'!I. __~_'------'-__'_'_r----"'_""""""""''''''''''''''~~

o 10 ;0 :l<I .0 &I ~ 10 10 jQ 100

Di:;p!Jccmcnl Dis1al1Cc (1Jn)

FIgure 3.28, Estimated change in calf survival during June for thePorCtlpine caribou herd, 1985-2001. as a function 01 the dislance ofdisplacement of lhe annual calving ground and associatedconcentrated calving alea and calving siles. Upper and lower dashedlines indicate 95% confidence inlervals on the mean effect.

11lell the empiricallllodcl was again used to predictc,llf survival for the displaced calving ground. 111edifference between tlle calf survival estimate for thedisplaced and observcd calving ground was cnlcul;ttedand a dmaset of 46 displacement distances and associatedchanges ill calf survival was gencr3ted for 3nalysi~.

111e model showed a significant (r = 0.47, P < 0.001)inverse relationship between displacement distnnce nndpredicted change in calf sllrvivnl (Fig. 3.28).

The simulations indicated that :1 substantial reductionill calf sun'ival during June would be expected under fulldevelopment of thc 1002 Area. Eighty-two percent ofobsen'ed calving distributions would ha\'e been displacedand the aver:lgc distancc of these displacements wouldhavc been 63 kill (r,mge l6~99 kill). 'nlis would haveyielded a net average effective displacement of 52 kill andan expected mean reduction in c;llf survival of 8.2% (SE= 0.71,~O.

It is remotely conceivable that calving caribou of thePorcupine caribou herd could select habitats tllat yieldedequivalent fnrage and predation risk after displacemcnt.Forage for lactating females of the Central Arctic herd,howcver, declined as the concentrated calving area in thedeveloped wne shifted to the south-southwest (Wolfe2000). This SllggcSI.~ that such compensatory habiwtuseby the Porcupine carihou herd would be unlikely if theircalving grounds were displaced by oil development.

BccausL' there was no empirical basis for changing theshape of the {)bscrved calving distributions, it W;l.~

impossible to estimate the magnitude of the effect ofconsidering the peripheral calving areas and calving sitesas missing data when they were displaced onto the ocean.Tile effect wa$ expected to be snl,lll. Arbitr:lrily assigningcalvin!!. site.~ that were displaced onto the ocean back ontothe Cl);stal plain and making no otlwr 'H.1justmcnts would

32 BIOLOGICAL SCIENCE REPORT USGSIBRD 2002-0001

lave increased displaced calf survival by only about O.6S-;'In average. lllis probably constituted the maximumlossible C£fect of treating areas and calving sites thait"crc displaced to the BC[lufort Sea as missing dat[l.

Stochastic simulation modeling (Walsh et al. 1995)ldicatcd that a 4.6% reduction in Porcupine caribou herdalf survival during June, all else held e(lual, would havecen sufficient to halt growth of the Porcupine c[lriboucrd during the best conditions observed to date. A lO-km\'erage displacement in our simulations would have beenJfficient !O bring the upper confidence itllerval on thelean effect below a 0% predicted change in calf survival'ig. 3.28). A mean displacement of 27 klll in our:odeled prediclion.~ would have been sufficient to reachiC threshold of 4.69(· Illean reduction in calf survivallfficielll 10 halt growth of the Porcupine caribou herdlder best observed growth conditions 10 date. This lattervel of displacement could occur well bdorc full·.velopmellt of the 1002 Area.

The estimated effect of displacement of the Porcupincribou hen! Oil c[llf survival during June was·nserv:ltive for several reasons. First, we used thenservativc estimate of a 4 km displ[lcelllent I)fJlcen1rated calving areas from infrastructure (Callll:fOnal. 1992) vcrsus 7-8 kill (Wolfe 2000). Second, weIplaced tlle concentmted calving ;lreas parallel to the'aufon Sea coastline lll\ls maimaining calving:tributions on the bes't remaining coastal plain habiWI:I minimizing displ[lcement into the foothills where~dation would be expected to increase calf lllortality.lally, relatively low density calving was allowed to:rlap developed [lreas, as has been observed for thencent Central Arctic herd (Wolfe 2000, Lawhead ;lIldchard 20(1).Because the a;;sumptiolls were conservativc, IlICults were conservativc. Su!lstanti[ll (10 to 27 km)placemeJlt of concentmtcd calving areas and associatedlUal calving grounds and calving sites of thc Porcupincibou herd i.~ likely to negatively affect calf surviv;11ing JUlle. At tIle upper cnd of this range ofJl[lcement (27 km), recovery of the herd from the[ellt decline (Fig. 3.8) would be unlikely. Theseelusions are consistent with those found in the 1987II Legisl[ltivc Environlllentallmpaet Statement)Ugh et a!. 1987).The Ilorcupine caribou herd has demonstratcd,tantiainaillral variability in size alld demographyos. 3.5, 3.8, 3. IOIl"C). Because developllll~nt of the2 Area would take time, any cll"ec!.~ on the herd's'ornl:lncc may take (kcadeS to detec!. Reduced calfiralmay slow the rate of increaSe during positivc;es of the growth cun'e of the herd and increase theof decline during !he negativc phases of the herd'sI,th curve. The period of natur;ll cycles in herd size

may increase and the ;llllplitude of herd size Ill[lY beaffected.

The best empirical tool available for detectingpotential effects of development is the modeledrelationship between calf sun'ival ;lIld forage for femalesduring peak lactation demand (NDVI_62l) within theextent of t"[llving (Fig. 3.26). This model is independentof actual annual calVing ground location and encompassesa ne[lr full cycle of herd size [IS well as subsl[l/lti[llvariation in hemispheric we;lther patterns (Fig. 3.5) andvarbtion in calving ground location (Fig. 3.13).

With industrial development, if observed c;tlf survivalfalls below the lower 95{;~;' confidence limit Oil thepredicted observations from this Illodel (Fig. 3.26), or if aparallel pattern of calf surviv:ll yields a significantlylower inten;ept term, then an el"fect of development oncalf sun'iv[l! would be indicated.

Individual observations til[lt fall below the lowerconfidence limit and which call be satisfaelorily explainedby exceptional environmental characteristics (e.g., carry­over effects of ncar-catastrophic conditions in 1992 to1993 after eruption of I\f\lunt Pinatubo) (Fig. 3.26) neednot be considered evidence for effects of developmellt oncalf survival. A pattem of obsen'ed calf survival belowthe lower confidencc limit would be cause for cOllcen!.

Statistic[llme!hods for making these types of decisionsare currently in development (Rexstad and Debevec2001). l1tis assessmcnt will require continued intensivecal ving ground sun'eys and calf sun'ival estirn:l1es.

Conclusions

Our research has shown that the Porcupinc caribouherd has signific:lI1t annual variance in calving groundlocation (Fig. 3.13), f[lces annual vari;lIlce in habitatconditions, selects areas willI abulldall! high qualityforage for calving. has incre;lsed MIn-ivaI of calves bornin the concentrated calving areas, alld shows a con'clationbetween calf survival and both forage for females duringpeak lactadon ant! predation risk in the annual calvinggrounds. All this implies that unrestricted ;Jccess to annualcalving grounds and concentrated calving areasmaximized perfoonallce of lactating Porcupine caribouherd females and their c[llves. Becausc the Porcupinee:lriboll herd has shown limited c;lpaclty for growth, frel:access to calving ground habitats lIlay have compensatedfor less than oplimal wintering h;llJitats.

Location of the conceillrated cal ving areas during thepas! 19 years (1933-2001) is the best estimatc of thc arcathat has provided the highest qualit)' calving habitat forfemales and their calves. Calf s\lfvival within !heaggregate eXlem of conccntrated c;l1ving arcas has beenhigher (han fllr cal res born in [lreas never used as al:OIlCelllralcd calving area (83.8'_';~ vs.7J.9'/;', respectivdy,

ARCTIC REFUGE COASTAL PLAIN 'TERRESTRIAL WILDLIFE RESEARCH SU!'IHvlARIES :n

"CJ.."""'" 1111.'0' .-.",,,,,,,. , ..... ,

[]""""." .."., '" '<*':''''''''''0 ..,,,..,. ,..._"-OC'......O, .... '"""'0 "OCOO>< 1.,,:00"'"

,..,url C}.." ..... ..".,."LOOI.. ,_..'.."'00.1"'...0'0"''''.....'

\, ,. ,"

Figure 3.29. Aggregate extenl of annual calving (light green shading)and aggregate exlenl of cnncentrated calving (dark green shading) forthe Porcupine cariboU herd, 1983·2001. The deformedlundelormedgeological boundary is discussed in USGS Fact Sheet FS-028-01(U,S. Geological Survey 2001),

19l-:3-1994, p:::: 0.026). 'nlUS, the aggregate extel1! of allobserved concentrated calving areas (Fig. 3.29) identifiesthe most valuable portion of the extent of calving in terms

of calf survival during June.Our [llOdel prediction of ,I reduction in calf survival

when calving, g,rounds were displaced supports theconcept that caribou made a critical "decision" in locatingtheir annual calving grounds within the extent of calving.19:-:3-2001. It appears that actual calving ground locationmaximized June calf survival given the habitat conditionswithin the extent of calving for a given year.

Weight-g,lin of calves provided further evidence for',he import:lnee of unrestricted localion of ,1I11lual calvinggrounds. The lack of a relationship IJl?tween calf wcight­~~;lill and habitat usc within annual calving groundss\I!:\!ests that weight-gain was optimized by selection ofth~ ~Innu<ll calving gr~unds. particularly during the first 3

weeks of life.Cornparalive growth of captive and wild P{)[cuJline

caribou herd (:alves (Parker et al. 1(90) has shown thatwild Porcupine caribou herd calves attain their maximuill~~(:!~e{k potenti':.ll for d:lily wei ght-gain during, early- toill:d·lactation (Gerhart et ,Ii. 1(96). Therefore unrestrictcdc:<.:!cction of the annual cal ving ground may optimize',·.'eight-gain of calves for a year- The matching r:lllk ordasIlrt'-lDVI_621 in the annual calving grounds and calfwc'\ghtS:H 3 weeks of age. 1992-1994, support thiscilil<.:ept.

UnreStricted selection of annu;ll calving grounds likelyh;.;j significant implications for the p:lrturient females as',','el1 ;IS for their calves The ll1atchin~ rank orders of I)NDVI_621 within mlllual calving grounds, 2) parturient

femalc weights. and 3) parturient female body conditionsC(Jres during peak lactation. 1992-1994, suggcstsubstantial contribution of the calving ground 10parturielll females' lllllritional status. Bccause fall wcightsof panurient females influenec their probability ofconception {Camcron et ai. 1993, C:lmeron and va Hoef1994, Russell et ;11. 1(98). c:tlving ground habitats maycontribute 10 parturition rmcs in the following year.

Petroleum developmcnt will most likely result inrestricting the location of concentratcd calving areas.c:llving sites, ,lIld annual c:llving grounds. Expectedcffect~ that could be observed include reduced surviv,ll ofcalves during June. reduced weight ,lIld condition ofparturient females :md reduced weight of calves in lateJune, ,md. potentially, reduced weight and reduccdprobability of conccption for parturient fcmales inthc

fall.Whether these factors :lrc additive to annual

performance or arc compensated on wimer range willdetermine the nct value of the annual c:llving grounds 10

herd performance. Detcnnining the additivcJcompensatory nature of allnual calving ground value,through field and simulation studies, should be the fir~t

research priority in future workStill unclear is the cause of the decline of the

Porcupine caribou herd (Fig. 3.8) during a period whencalving ground habitat conditions were favorable as aresult of summa warnling. Increased winter JllOrtal ity wasimplicated by the herd decline because sub-adult and:ldult mortality on the c;llving ground has beenineonscquenti,l1 (Fancy et:ll. 1994, Walsh et al. 1995),,wd parturition rate and calf survival during June hasremained high during the decline.

Possible mechanisms for this suspected increasc inoff-calving-ground mortality include: I) reducedIOIl"evitv or adult females as a result of the cumulative

~ -cncrgetic costs of persistem high parturition and calfsurvival during climate warming. 2) incre;lscd energeticcosts of insect harassment:ls the climate has warmed, 3)reduced availability of winter forage or other adverseeffects associated with increasing frequency of freeze­thaw cvents. 4) thc herd exceeded forage carryingcapacity of winter range, or 5) an increase in some formof predation (human or natural) on tlte winter range.

Increased frequency of spring and fall icing events onnon-cal \' ing hal)lt;Il.~ of the Porcupine caribou herd (Figs.3.71l./J) supportS the third hypothesis and lIlay beimplicatcd in the fifth hypothesis (increased predationmortality). Increased frequency of icing was not evidenton the non-calving ranges of other Alaska barrcn~gr()und

caribou herds that have not declined significantly duringthe 1990s (Central Arctic herd, Teshekpuk Lake herd,Western Arctic herd). Testing the remaining hypotheseswill require subslalllial additional fieldwork.

14 BIOLOGICAL SCIENCE REPORT USGSfBRD 2002-0001

In summary, 4 research-based ecological argumentsIdicatc that the Porcupine caribou herd may beanicularly sensitive to development within the 1002;::ll1ion of the calving ground:

tva' nnJiill({;vjrv Q(rht' Pnrrwi!lf Cllri!JQII herd -111ePorcupine cnribou herd has had the lowest c:lpacityfor growth among Alaska barren-ground herds(Porcupine caribou herd = 4.9%, Central Arcticherd =10.8%, Teshekpuk Lake herd =13'70,Western Arctic herd = 9.5%) :lnd is the onlybarren-ground herd in Alasb known to be indecline throughout the 1990s. lllis low growth rate(Fig. 3.9) indicMes that the Porcupine caribou herdhas kss capac it}' to accommodate anthropogenic,biological, and nbiotic stresses than other Alaskabarren-ground herds. Any absolute effect ofdevelopment would be expected to have a largerrelative effecl on the Porcupine cariboll hcrd thanon the olhcr herds. For examplc, an approximatc4,6t,1> rcduction in calf survival, all elsc held equal,would be cnough to prcvcnt Porcupine caribouherd growth under the best conditions obserwd todate (Walsh et al. 1995) or prevent recovery frolllthe current decline. A similar reduction in calfsurvival, all c1sc held equal, for other Alaskabnrrell-ground herds, howcvcr, would not bcsufficient to arrest their growtll.

lk/lltl1l Hr<II,'<Giliti n(COIlccnrT(I(j'd wid/! R<l.liJlSJ.Ji

thf Cel/tral ATc1jc...£G..!i.lli.lliJJ.cr!i..all'(/v (mm

n,'tm!lICIIl dD:.d.J.)j!III('1l1 jlltra.WTIICWT('S - It isassullled thm the Porcupine caribou herd caribouwill avoid roads and pipelincs during calving in amanner similar to the Central Arctic herd ifdevelopmcnt of the 1002 Area occurs. Avoidanceof petrolculll development infrast ructure bypanuricnt caribou during the first few weeks of thelives of calves is the most consistently observcdhelmvioral response of caribou to devclopmcnt.

l{wk (J(birlhJ/lillilv a!I"l"Iuj«' coli';'!!: }wbiliU­Calving areas in Canada and away from thc Alaskacoastal plain were used only whcllthe ArcticRefuge l'{lastal plain, including the 1002 Area,were unavailable due to late snowmelt. Diet qualityon the Canadian portiolls of the calving groundwas substantially lower than on the Arctic Refuge("oastal plain and i002 ponio!1s oftlle calvingground. \Vhen snow cover reduced access byfemales to the Arctic Refuge coastal plaill and1002 Area for calving, calf survival during lUllewas 19~;' lower than when they could calve on theArctic Refuge cl):lstal plain and 1002 Area.

Strong link between ca/(suITjva{ nnd free wQl'emenrn((ema1cr - TIle location of the llilnual calvinggrounds and concentrated calving areas wasvariable among years in response to variablehabitat conditions and was oftcn coincident withthe 1002 Area, Empirical relationships betweencalf survival, forage nv:lilable to fcmales in theannual calving grounds, and predation risk derivedfrolll 17 years of ecologiclll data predict that lunccalf survival for the Porcupine cnribou herd willdecline if the calving grounds arc displaced, andthat the effect will increase with displacementdisl:lncc. This prediction (Fig. 3.28) is a functionof displacement: I) reducing accesS 10 the highestquality habitats for f(mlging and 2) increasingexposure to risk of mortality from predation duringcalving (lirst 3 weeks of June).

References

Akaikc, H. 1973. Inforlllntion theory, an e.'tension of thenl:l.,imumlikclihood principle. P~ge.s 267-281 in B. N.Petrol' 3.nt! F. Csaki, editors. Second InternationalSymposium on Infonll~tiol1 Theory. Akademiai Kaidi,Budapest, Hungary.

Allaye-Chan, A. C. 1991. Physiological and ecologic;lldeterminanlS of nUlrienl panitioning in caribou and reindeer.Dissenation, University of Alaska, Fairbanks, Alaska, USA.

Baglio, J. V, and E. \V. Holroyd, Ill. 1989. Methods foroperalional snow cover area mapping using the adv:lIlcedvery high resolUlion r.ldiometcr - 5;1ll Juan Mountain leststudy. Research Technical Report, U.S. Geological SurveyEROS Dala Cetllcr, Sioux Falls, S{)ulh Dakola, USA.

Bergerud, A. T. 1980. A review of the popubtion dynamics ofcaribou and wiM reindeer in North America. Pages 556·581ill E. Reimers, E. G~arc, and S. Skjenncbcrg, edilors.Proceedings of the Second Itllcrn:ltional Reindeer/CaribouSymposium, Roms, Norway. 1979. Dircc{()ralct for yill ogfc'rskv,mnslisk, Trondhcim. N'lrway.

__, and R. E. Page. 1987. Displacemclll and dispersion ofp;muricnl caribou at calving as ;lntipn:dator taclics.Canadian Journal of Z.oology 65: 1597-1606.

Billings, W. D., :lIld H. A. 1\loollcy. 1968. The ecology of ~rclic

and alpine pl:lnts. Biological Review 43:481-529.Bliss. L. C. 1971. Arctic ;\lld alpine life cycles. Annual Review

of Ecology and S)"Stematics 2:405·438.Brown, R. D., and R. O. Braaten. 1<)98. Spatial and tcmllilr;J]

variability of monthly Canadian snow depths, 19·16-1995.Atmosphere and Ocean 36:37-5'1.

Cameron, R. D. ]<JS3.lssue: c;triboll ;llld pdro1ctlilldevelopmcnt in arclic Alaska. Arclic. 36:227-231.

__~,,_, D. 1. Rc'cd, J. R. Dau, and W. 1'. Smith. 19<)2.RC'distrib\ltion of caribou in response to oil fieldclevel 0plllenls OJ! lh<~ ,·\relic Slope of Alasb. Arctic 45:331\­3012.

,W. T. Smith. S. G. Faun', K. L. Gerhart, :lIld R. G.White.-~ .

1993. Calving su<:<:e,s of fcmak caribou in relation 10 bodyweight. Calladian Journal of Zoology 71 :·n;o··1l:l6.

.~,ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 35

III

__•and J. Vcr Hoef. 1994. Predicting parturition rale ofcaribou from autumn body mass. Journal of Wildlifet-.hnagclllcnl 58:674-678.

__, and K. R. Whiucn. 1979. Seasonallllovemenls andsexual segregation of caribou dctermined by aerial survey.JOtInl:l1 of Wildlife ~lanagemenl 43:626-633.

__, and __,1980. Nutrient dynamics of caribou forageon Alaska's Arctic Slope. P;lges 159-166 in E. Reimers, E.Gnare. and S. Skjcnllcbcrg, editors. Proceedings of theSecond Illlcrnalional Reindecr/Carihou Symposium, Roros,Norway. 1979, Directorate[ for vilt og fcrskv:ull1sfisk,Trolldhcirn, Norway.

Chan-IvIcLcod, A. C. A., R. G. White and D. F. Holleman. 1994.Effects of protein ;lnd energy intake, bod)' condition, and~eason on ll\llnent panilioning and milk production incaribOll and reindeer. Canadian ]ourn:ll ()f Zoology 72:~n8­

947.__, __, and D. E. Russell. 1999. Comparative body

composition strategies of breeding :lnd nOll-breeding femalecaribou. Canadian Journal of Zoology 77: 190 1-1907.

Chapman, W. L, and J. E. W:llsh. 1993. Recent variations of seaicc ;Ult! air temperatures in high latitudes. Bulletin of theAmerican Meteorological SocielY 74:33-47.

Clough, N. K., P. C. Patlon, and A. C. Christensen, editors.1987. Arctic National Wildlife Refuge, Alaska. coastal plainresource :lsseSsmcnt - report and recomll1end;ltion to theCongress of the United States and final legislativecnviroumelllal impacl statement. U.S. Fish and WildlifeSCf'.'ice, U.S. Geologie-II Survey, and Burcau of LandManagement, WaShington D.C., USA.

Cronin. M. A., S. C. Amstrup, G. M. DUfner, L. E. Nod, T. L.McDonald, and W. B. Ballard. 1998. Caribou distribUlionduring the post.calving period in rebtion to infr;lstrucluTe inthe I'mdhoe Bay oill1eld, Alaska. Arctic 51 :85-93.

..,, , W. B. Ballard. J. D. Bryan, B. J. Pierson, and J. D.McKendrick. 1998. Northern Alasb oil fields ;md caribou: aC()11l111elllary. Biolol~ic:l1 Cmlscrvalion 83:195·208.

Curatolo, J. A., alHJ S. M. Murphy. 1986. The effects ofpipdines, mads, and tramc on the movel1letl!s of caribou.Rimgifcr lilfilndlls. Canadian Fiehl·Natur;llist !OO:218·224.

Dau, J. R., and R. D. Cameron. 1%6. Effects of a HJ:J.d systemon caribou oJistriblllion during calving. R:lI1gifer. SpecialIssue 1:95-101.

Duquellc, I.. S. 1984. Pnll,~ms of activity and their implications tolhc ,'nergy budget of migr;lting caribou. Thesis. University ofAlasb, Fairbanks, Abska. USA.

Fancy, S, G.. L. F. Panko K. R. Whitten. and W. L Regclin.1989. Seasonal mO\'eillcnts of caribnu in arclic Alaska asoJctennined hy .~atellite. Canadian Journal of Zoology67:644·650.

___' :Illd R. G, While. 1987. Energy "xjlenditllfcs for10coillolinll hy harren ground carihou. C,llladbn JOllfllal ofZoology 65:122· 128.

__. and K. R. Whitten. 1')<j I. Selection of (alving sit,'s bythe I"lr,'upine C:lri\Joll herd. Canadian Journal of Znology{i9: 1736-1743.

__' __. ~llld D. E. Russell. 19')4. Demography of thePorcupine c~lri1JOIl herd. I9K~·1 992. Canadian Journal ofZoology n:R·lO·8·j6.

Fryxell, J. M. 1991. Forage quality and ~lggregati()a by largeherlliyores. ;\lllcri,':l11 N:llUralisl IJS:-17R·4i)l\.

__.1995. Aggregation and migration by grazing ungulatesin rel:ttion to resources and predators. Pages 257-273 in A.R. E. Sinclair and P. Areese, editors. Screngeti II: dyn;lmics,management ;lnd conservation of an ecosystem. Unh'ersityof Chicago Press, Chicago, JIIinois, USA.

Gtrllal1, K. L., R. G. White, R. D. Cameron, and D. E. Russell.1996. Body composition ~ll1d nutrient rcscn'cs of arcticcaribou. C:tnadian Joumal of Zoology 74:136-146.

Groisman, P. A., T. R. Karl, and R, W. Knight. 1994. Observedimpact of snow cover on the heat balance and rise ofcOlllillenlal spring tempcralUres. Science 263:198-200.

Gro~s, J. E., L. A. Shipley, N. T. Hobbs, D. E. Spalingcr, and B.A. WU!lder. 1993. Funelional response of herbivores infood-concentrated patches - tests of a Illechanistic model.Ecology 74:778-791.

Hamilton, W. D. 1971. Geometry for the selfish herd. Journal of11lCorelical Biology 31 :295-311,

Hare, S. R., and N. J.lIfantu:l. 2000. Empirical evidence forNorth Pacific regime shifts in 1977 and 1989. Progress inOceanography 47: 103·145.

Ilaukioja, E., and R. Salovaara. 1978. Summer weight ofreindeer (Rilllgifa laralldll.l) calves and its irnponance fortheir future survival. Repans of the Kevo SubarcticResearch Stnlio!l 14: 1-4.

Ilellc, T., and L. Tnn·ainen. 1984. Effecls of insect harassmenton weight g;lin and survi\'~t1 in reindeer calves. Rangifer4:24-27.

Holben, B. N. 1986. Characleristics of maximum-valuecQmpo~ite images fWllllemporal AVHRR data. IntemationalJoumal of Remote Sensing 7: 1417.

Hope, A. S .. J. S. Kimball, and D. A. Stow. 1993.111Crelationship between tussock tundra spectral renectanccpropenics nnd biomass and vegetation composition.Intemational Journal of HeT110le Sensing 14: 1861-1 874.

Houghton, J. T., L. G. Mcira Fihlo. B. A. Callender, N. Ham, A.Kanenherg, and K.1I1askell. 1996. Clim:ltechange 1995:the science of climate change. C:llllbridge University Press,Cambridge. England, United Kingdom.

Johnson. D. H. 1980. The comparison of usage and availabilitymeasurements for cvalualing resource preference. Ecology61 :65-71.

Jorgenson. J. c., P. E. Joria, T. R. 1\IcCabe, B. R. Reitz, M. K.Raynolds,lIl. Emers, M. A. Wilm.~. 1994. Users guide forthe land-cover ll\;IP of the coastal plain of the ArcticNational Wildlife Refuge. U.S. Fish and Wildlife Service,Fairbanks, Alaska, USA.

Kelleyhousc, H. A. 2m)!. Calving ground selection and fidelity:Westem Arctic aml Teshekpuk Lake herds. 'nlCsis,University of A!:lska. Fairbanks. Alaska, USA.

KUTOpat, P. j, 1984. Foraging behavior of caribou on a calvinggrOlllld in Nonh\\'e.'itcrn Alaska. Thesis. University ofAI:I'ik:l. fairbanb. :\Ia~k:t, USA.

L"lngvatn. R., and T :\. Hanley. 1993. Feeding-patch choice hyred dccr in rebtioll to foraging eflicicncy - an e.\]Jerimelll.Oecologia 95: 16-l-170.

Lawhead. B. E. I9Sfi. Diwihutioll ;md movcments of C~ll1rnl

Ardic herd cmibou during the calving ruHl insect Se;lS(}[l..,.l'roecetlinp of the Third North Americ'an CaribouWorkshop. Alaska Depallillent of Fish and Game. WildlifeTechnical Bulletin S:8·13.

36 BiOLOGICAL SCIENCE REPORT USGS/BRD 2002-0001

__• L C Byrne. :lnd C B. Johnson. 1993. CaribouSynlhesis: 1987-90 Endieoll environmentalmoniloringprogram. Ullpublished report sponsored by U.S. ArmyCorps of Engineers, Alaska District, Anchorage. Alaska,USA.

___, and A. K. Prichard, 2001. SUf\'eys of caribou :llldmuskoxen in lhe Kuparuk-ColvilJc region, Al:lska, 2000.Final Rl'port for Phillips Alaska. Inc., prl'parcd by ADR.Inc., FJirbanks. Alaska. USA.

tllalingreau,1. P.• and A. S. Ilc!w;lfd. 1992. Scale consideralionsin vegelation monitoring using AVHRR d:lla. Intern:llionalJournal of Remole Sensing 13:2289-2307.

?\larkon. C. J.,I\l. D. Fleming. :Illd E. F. Dinian. 1995.Characterislics of vegelJtion phenology over the Alaskanlandscape using AVHRR lime-series data. Polar Hewnl31:179-190.

Maxwell. B. 1996. Recelll climale paltcrns in lhe Arctic. Pages21·46 in W. C. Deche!, T. Callaghan, T. Gilmanov, J. I.Hollen, D. M;l.,well, U. Molau. and B. Sveinbjornsson,edilors. Global change 'lIld arclic terreslrial ecosyslems.Ecological Sludies 124, Springer Verlag, New York, USA.

Mielke, P. w., and K. 1. Berry. 1982. An eXlended class ofpenmJlation leclmiques for malched pairs. Communicalionsin Slalislics: Theory and Melhods II: 1197-1207.

Minnis, P.• E. F. Harrison, L L Slowe. G. G. Gibson. F. M.Denn. D. R. Doelling, and W. L. Smilh, Jr. 1993. Radiativedimate forcing by the MOlllll Pinatubo enlption. Science259:1411-1415.

Murph)', S.M.. 'lUd J. A. Cur:uolo. 1987. AClivity budgets :llld1l10velllezll rales of euiboll encoulllenng pipelines, roads,and lraffic in Iwrthem A[;lsI:<I. Canadian Journal of Zoology65:2483-2·190.

__, and B. E. Lawhead. 2()()]. CarilJllll.l'ages 59·S·1 in J. C.Truett and S, R. Johnson, editors. 'flle natural history of anArctic oil field: devclopmelll and lhe biOla. ,\cademic PresS,San Diego. California. USA.

Mym:ni. R. B.. C. J. Tuckcr, G. Asrar. and C. D. Keeling. 1995.Inlerannual varialions in sale IIi Ie-based vegetation indexdala from 19S I 10 1991. Joumal of Geoph)'sical Research06 103:61·15·6160.

Mynelli, R. B.,C. D. Keeling, C. J. TUcker. A. Asrar, and R. R.Nelllani. 1997. Increased plant growth in the northern highlatitudes from 19R1to 1991. N:uure 386:WS-702.

Nellemann. C. and R. D. Camaon. 1996. Erfec1.~ of petroleUIlldevc!oplllelll on terrain preferences of calving caribou.Arclic ·19:23·21\.

__. alld __. 1998. CUJJlUI:ltive ill1pacl.~ of:m evolvingoil-field complex on the distrihlllion of calving carihou.Canadian Journal of Zoology 76: 1425·1430.

Nixon, W. A. C. 1990. Group dynamics and heha,'ior of thePorcupine carihou herd during the inseel season. 'illes is,tJnin~rsity of Alaska, Fairbanks, Alaska, USA.

O'Neill, R V.. and A. W. King. 1998. Homage lO Sl. Micllacl:or, why arc lhere so many hooks on sC:lle.Page's 3·16 ill D.1.. l'c1ersoll and V. T. P:lIt",r, ",ditors. Ecological scale:theory and applications. Columbia Universily !'r",ss. NewYork, New York. USA.

l'~rker, K. L., R G. While. M. P. CJillillg!l;II11. and D. F.Holleman. 1990. Comparison of energy Illelabolism inrclation /<) daily ;,cth'il)' a)ld milk con.'iumption hy caribou;Ind rlltlsko,\ lleonat<'~. Canadian JOlllllal of Zoology 68: lO6­11·1.

Paruelo, J. 1'11 .• B. E. Epslein. W. K. Laucnrolh, and I. C. Burke,1997. ANPP eslimales for NDV! for lhe ccnlfal grasslandregion of lhe Uniletl Slales. Ecology 78:953-958.

Pollard, R. B., W. B. Ballard. L E, Noel, antl M, A. Cronin.1996. Surnmer dislribution of caribou, Rangifcr tarmrdusgran/i, in lhe area of lhe Prudhoe Bay oil field, Alaska,1990-1994. Canadian Fic!tl-Naturalisl 110: 659-674_

POSI. E., and N. C. Stenselh, 1999. Climalic variabilily. plan Iphenology, and nonhern ungulales. Ecology 80: 1322-1339.

__, __. R. Lang..:lIn, and J. Fromentin. 1997. Globalclimate change and phenolypic varialion alllong red deercol IOns. Proceedings of lhe Royal Society of London.Series B 264:1317-1324.

Rexslad, E. A .. and E. Debe\'ec. 2001.1I1lerislic melric forecological I!lonil\)ring. PapCf presentcd atlhe AnnualMeeling ofllie Wildlife Sociely. 28 Seplember, 2001, Reno.Nevada, USA.

Roffe, T. J. 1993. Perinalal IlIonality in cariholl from thePorcupine herd, Alaska. Joum:l1 of Wildlife Diseases29:295·303.

Hussell. D. E., K. L. Gerhan, R. G. While, and D. van deWelering. 1998. DeleClion of early pregnancy in caribou:evidence for lael:llional infertilily. Journal of WildlifeManagemelll 62:1066-1075.

__, A. M. Marlell. and W. A. C. Nixon. 1993. Rangeecology of lhe Porcupine caribou herd in Canada. RangiferSpecial Issue 8.

__, and R. G. White. 1998. Surviving in the north· aeonceplUal model ofreprodUCliYe stralegies in arcliccarihou. Proceedings of lhe Eighth North American CaribouWorkshop, 23-25 April 1998. Whilehorse, Yukon Terrilory.Canada.

Sakamoto, Y., M. bhiguro, and G. Kilagawa. 1986. Akaikeinformation crireria slalistics. KTK Scienlit1c. Tokyo, Japan.

Seaman, D. E., B. Griffith. anl! R. A. Powell. 1'J9S.KERNELHR: a program for estimaling animal home ranges.Wildlif~ Socict)' Bulletin 26:95-100.

__, J. J. 1....lilIspaugh. D. J. Kemohan. G. C. Brundige. K. J,Raedcke, :md R. A. Gilzen. 1999. Effects of sample size onkemcl home range eSlilll,lleS. JoullIal of WildlifeIvlall;\gerllent 6):739,747.

__.•md R. A. Powell. 1996. An eValUalion of lhe a....curacyof kemcl densily e~til1\;u(lrs for home r:luge analysis.Ecology 77:2075-2085.

Serrele, M. c., J. E. Walsh, F. S. Chapin Ill, T. OSlerkamp. 1\1.Dyurgcro\', V. ROlllano\'sky. W. C. OecllCl. J. !\lori~on, T.Zh:Ulg, and R. G. Barry, 2000. Observation:ll el'idenee of recenlchange in the Jlorthern high latitude envimlllnenl. ClimalicChange .16: 159,207.

Shiple)". L. A., and D. E. Spalinger. 191J2. Mechanics ofbrowsing in den~e food palches - ,~ff,~cts of plant and animalmorphology on intake rate. C:U1:tdian Journal of Z~)()logy

70:1743-1752.Shippen, M. M .. D. A. Walk.:r. N. A. A\lerbach, :lIld D. E.

Lewis. 1995. Biollla."s :lIllllcaf-area index maps derivedfrom SPOT im:lges for T{)olik Lake and Illlnaval Creekar<"IS. Alaska. Polar Record 31: 1-/7-15·1.

Si!n:rlll:,n, U. W. 1'..l1l6, Density eslilll:uiou for Slalislie~ andd:lla analysis. Chapman and Hall. London. Engl,l1Id, UnitedKingdoJl\.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 37

I

I

Smith, W. 1'., and R. D. Cameron. [983. Responses of caribou[0 industrial development on Alaska's Arclic slope. ActaZoologica Fellic:! 175:43-45.

Spalingcr, D. E, T. A. Hanley, and C. 1'. Robbins. 1988.Analysis of the functional response in foraging in the Sitkablack tailed dca. Ecology 69: 1166-1 J75.

Sp:lrks, D. R., and J. C. Malcchck. 1968. Estimating ,l<:rccnlagc

dry weight in diets using a microscopic [cchniljlle. Journalof Range M:ll1agcl11cnt 21 :264-265.

Stone, R. S., J.lt Key. and E. G. DUllOlL 1993. Propcnics noudecay uf slralOsphcric aerosols in the Arctic following the1991 eruptions of Mount I'in:llubo, Geophysical ResearchLcHers 20:2359-2362.

lllOlllpWll, D. C., and K. H. l>!cCourl. 1981. Season;ll diets ofthe Porcupine Caribou herd. American l\lidl;md Nnturalist105:70·76.

__, ,wd 1. M. Wallace. 1998. The Arctic O~eiJlation

.~j£nalUrc in wintertime geopotclllial height and tempnalllrefield,. Geophysical Resean::h Leners 25: 1297-1300.

__, ;wd __. 2001. Regional climate impacts of theNonhem Hemispherc Annular Mode and associ:lled climatetrends. Science 293:85-89.

Toupin, B., 1. Huot, MlIJ M. Manseau. 1996. Effect of insectharassment on the behaviour of the Riviere George caribou.Arctic 49:375·382.

Trudell, 1., and R. G. White. 1981. The effect of forage structureand aV;lilability on food inl;lke, biting rate, bite site, ,mddaily eating time of reindeer. Journal of Applied Ecology18:63-81.

Tucker, C.l. 1979. Red and photographic infrared linearcombin,ltions for monitoring vegetation. Remote Sensing ofthe Ellvironment 8: 127-150.

___,J. Y. Fling. C. D. Keeling, amI R. H. G:Ulllllon. 19l16.Relationship bCllwcn ;llmospheric CO. variations and asatellite·derived vegetation index. Nall'lfe 319: I 95-199.

__, and P.l. Sellars. 1986. S,llellite reIllOte seming ofprimary production. Intemationallnumal of RemoteSensing 7:1395-1416.

Tussing. A. R., ,md S. Haley. 1999. Dr:linage pierces ..\NWR inAlJska study scenario. Oil and Gas JOllmaI97:71-X·l.

U.S. Geolog,ical SUn'ey. 200 I. .·\rcljc National Wildlifc Refuge,1002 Area. petroleum asse.~Slllent, 1993, induding economicanalysis. U.S. Geological Survey Fact Sheet FS·02g-01(Supcrcedcs FS-0·1O-98) April 2001.

Walsh. N. E., S. G. Fallcy, 'f. R. McCabe, and L. F. 1':l/lk. 1992.Hahi!;!t usc by the Porcupine caribou herd durillg predictedinsect h:lr'lssment.loumal of Wildlife Management 56:,165·473.

__• B. Griffith, and T. R. McCabe. 1995. Evaluating gmwthof the Porcupine caribou herd using a stochastic Jllodcl.lourn:llof Wildlife Matl:lgemellt 59:262-272.

__' T. R.l\IcCabe.l. M. Welker. and A. N.P:lrsons. 1997.Expc-rimelltal manipulations of snnw-tkpt!l: effects onnutri,-nt conlc-nt df ,',\fibOli (llf:lg<~. Ghlb:l! Change Biolo,:y 3(Supplement I): 158-1 (i·I.

White, R.G, 1983. For;lgin~ p;lttems and thdr lJmhiplicr eJfccls')Il productivity of northern ungulall's. Oikos ·10:377 ·384.

_-.",.._.1992. Nutritioll in relation to season, lac-tat ion, andgrowth of north temperate deer. P:lges ·107··117 ill R. D.Brown, editor. 11lc bi,)logy of del-r, Srringn·V,-rla!!. NewYork, Nell' York, US.-\.

__, F. L. 13unndl. E. Gaarc, T. Skogland, and B. Huben.1981. Ungul:ltes on arctic ranges. Pages 397-483 in L.C.Bliss, O. w. Ik:ll, and 1. J. Moore, editors. Tundraceo$ystems: a compar:lti\'e analysis. IrHemational BiologicalProgram 25. Camhridge Universit)· Press, Cambridge.England. United Kingdom.

__, ,lf1t! J. R. Luick. 198-1. Plasticity and constraints in the!:lctation:J! strategy of reindeer and caribou. Symposium ofthc Zoologic:lI Society of London 51:215-232.

.._._' B. R. ·nlOmson. T. Sk!)gl:md, S. J. Person. D. E. Russell,D. f. Holleman. :llld 1. R. Luick. 1975. Ecology of caribouat !'mdhoc Bay. Alaska. Pages 151-201. ill J. Drown, editor.Ecolo,gie;ll investigations df the tundra biome in thePrudhoe B,ly regioll, Alaska. Biologic:ll P;l]Jers. Universityof Abska, Special Report 2.

___ , and 1. Trudell. 19S01l. Patterns of herbivory and nutrientintake ofrcindecr grazing tundra vcgct'llion. Pagcs 180-195in E. Reimers, E. Gaare, allt! S. Skjcnncbcrg, editors.Pmceedings of the Second Intem:lliollal Reindeer/CaribouSymposium. Roros, Norway, 1979. Dire.ctoratet for .. ilt osferskvannsfisk, Trondheim, Norway.

__, and __. 198Gb. H;lhitat preference and forageconsumption by reindeer and caribou ncar Atkasook,Alaska. Arctic and Alpine Research 12:511-529.

Whitten. K. R., and R. D. Camcron. 1983. Movements ofcol bred caribou, Rilngifa {amlltllls, in relation to petroleumdevelopment Qutlle Arclic slope of Alaska. Canadian Field­Naturalist 97: 143-1·16.

__, and __. 1985. Distributiol1 of caribou calving inrelation to the J'mdhoc Bay oil field. Pages. 35-39 ill A. M.Mancil and D.E. Russell, editors. Caribou and humanactivity: proceedings of the First Nor1h Americ-:lll CaribouWorkshop. Whiteholse. Yukon Territory. 1983. CanadianWildlife Service Special Publicatillils. Ott:lwa, Ont;lrio,Canada.

___ , G. W. Garner, F. 1. Maua, and R. B. Harris. 1992.Prodnctivity ;md early calf survi\';ll of die Poreupinc c:lribouherd. 10llnl:ll of Wildlife M;uwgemellt 56;201-212.

Wiens, J. A. 1989. Spalinl scaling in ecology. FunctionalEcology 3:31:\5-397.

Wilmshurst, J. allll 1. Fryxcll. 1995. Patch seleclion by red deerin lc!atiDn to energy and protein illtakc: ,l re-evahJation ofLang,,"atn ;Illd Hanley's (I <)93) results. Oecologia 104:297­]00.

Wolfe, S. A. 200n. Habitat selection by calving cariboll of theCell1ral ,\rctic herd. 1980-<)5. 11lesis, [Jnh'ersity of Alaska,Fairbanks..·\la.~ka. USA.

___, n. Griffith, and C. A. Gray Wolfe. 201JI). Re~]lonse ofreindeer and carihlJll to hllnlan acti,itics. Polar Research19:6:1·73.

Young. D. J.• and T. R. McCabe. 1997. Griuly be:u predntionratcs ,m cMih,lU ca!l'"s in Ilorlhe~,tern AI;I.,LI. JOllrnal ofWildlife ~l:mag<'lllent 61: I()56·1 066.

Zhou. 1... C. J. Tuckcr. I{. K. Kaufmann. D. SI'lyhack. N. V.Shnh;l!lol" :Iud R. B. lIfyll<'lli, 2001. Vari:ltion in northern,cgetation ;tctil'ity infcncd from sat,'llii,' data of vegetationilHJcx durin,; 1lJ811<l1<j<)l).lollrnal ()fGeophysic~1Rese:lr~'h-~\tmo~]llicws 1O(j (D 17):20069':200?;:l.

38 BIOLOGICAL SCIENCE REPORT USGS/BRD 2002·0001

O••~,_~~....5===='it? MILES

\,,\ALASKA

,

(Dauphine 1976, Thomas 1982, Reimers 1983, White1983, E10ranta and Niemincn 1986, Lenvik et al. 1988,Thomas and Kiliaan 1991) and survival (Haukioja andSal{l\'aara 1978, Rognmo et al. 1983, Skogland 1984,Eloranta and Nieminen 1986, Adamczewski ct al. 1987).

Those concerns, though justified in theory, lackedempirical support. With industrial devc!opment in arcticAI:lska virtually unprecedented, there was little basis forpredicting the extent nnd duration of habitat loss, muchless the secondary sllort- and long· term effects on thewell-being of a particular caribou hcrd.

Furthennore, despite a general acceptance that bodycondition and fecundity of the felllales arc functionallyrelated for reindeer and caribou, it seemcd unlikely thatany single model would apply to all subspecies ofRmrgila, and perhaps not c-vcn within a subspecies indifferent geographic regions. We therefore lacked acomplete underst:lJIding of the behavioral responses ofarctic c:lribouto industrial development. the manner inwhich access 10 habitats might be affected, and howcll,mges in habit:lt use might tr:lnslate into measur:lbleeffects on fecundity and herd growth rate.

SPINE ROAD

KUPARUKPIPELINE

MILNE PT,

N

OLiKTOK PT.

Section 4: The Central Arctic CaribouHerd

o 5 10 KILOMETERSEM~.~.'='==ll

From the mid-1970s through the mid-1980s, use ofcalving and summer habital~ by Central Arctic herdcaribou (Rangifa tarafldlls gmnti) declined ncarpetroleum development infr:lstructure on :\I:lska's arcticcO:lstal plain (Cameron et a!. 1979; Cameron and Whitten1980; Smith and Cameron 1983; \Vhillen and Cnllleron1983a, 1985; Dau and C:llllcron 1986).

With surface development continuing to expandwestward from the Prudhoe Bay petroleulll development:m::a (Fig. 4.1), concerns arose that the resultantcUlllulative losses of habit:!t would eventually reduceproductivity of the caribou herd. Specifically, reducednccess of adult females to preferred foraging areas lllightadversely affect growth and fallening (Elison et a1.1986;Clough ct:ll. 1987). in turn depressing calf production

Raymond D. Cameron, Walta T Smith, Robert G. Whitt',

and Brad Griffirh

FIgure 4.1. Petroleum development irlfrastruclure in the Prudhoe Bay and Kuparuk peiroieuill development areas, Alaska, showing primary andsecondary roads. pipelines. and gravel pads, 199~.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RE":iEARCH SUMMARIES 39

Our study addressed the following objectives: I)estimate variation in the size and productivity of theCentral Arctic hcrd; 2) cstimate changes in theJistriblllion and l1lovements of Central Arctic herdcaribou in rdation to thc oil ficld dcvelopment; 3)estimate thc relationships between body condition andrcproductivc performance of female Central Arctic herdcaribou; and 4) compare the body condition, reproductivesuccess, fUld offspring survival of females underdisturbance-free conditions (i.e., cast of theSagavanirktok River) with the stmus of those exposed topetroleum-relatcd development (Le., west of theSagavanirktok River).

Status of the Central Arctic Herd

Photocensus results indicate net growth of the CentmlArctic herd from 1978 through 2000 (Fig. 4.2). Within

Herd Size

§:1 •

•

x j~20 •

j lS-

•

<; ! •ci 10Z •

•5· r-,- -,-,1976 "" '''' "'"

,,,..""

Net Calf Production

'00 •• ••~oo • •• •U •• •~ 00..~ •• •• •• •, 70 •E•~ • •

"ro •

" i"• •

\S'82' '--'-T·-·'--'-··-~-'-'-'-'

1978 "" "'",,,..

""

Figure 4.2. Pholocensus estimates of lhe Cenlral Arctic caribou herd,1978-2000 [Whitlen and Cameron 1983b; Alaskii Department of Fishand Game (AOF&G) files] and net calf production based onobsel\'alions 01 radio-collared adull (i,e" sexually.mature} femaleslrom 10 June through 15 August (ADF&G files). Nole: Procuclivity dalanot adjusted for differences in sample sizes east and west of lheSagavanirktok River, Alaska,

that long-term trcnd, however, there was an abruptdecrc:lse from 1992 to 1995. This decrease coincided withcalf production estimates at or bclow approximatcly 70%.Steady growth thereafter was associated with productivityestimatcs consistcntly exceeding 70%.

Development-related Changes in Distribution

Since 1978, changes in the distribution of calvingcaribou :lssociated with the Kuparuk petrolcumdevelopment :lrC:l, wcst of Pmdhoe Bay (Fig. 4.1), havebeen quantified using stri]HranSecl surveys flown byhelicopter.

After construction of a road system ncar Mi]ne Point,mcan caribou abund:lnce declined by morc than two­thirds within 2 kill from :l road and was less thanexpected, overall, within 4 km; but nearly doubled 4-6 kmfrom roads (Fig. 4.3) (Cameron et al. 1992b). Prior toroad placcmcnt, caribou were found in a single, morc-or­less continuous concelltration roughly centered where thc1\-1ilne Point Road was subsequently built. AfterCOl1stmction of the road, a bimodal distribution withseparate concentrations east and wcst of the road W:lS

clearly apparent (Fig. 4.4) (Smith and Cameron 1992),indicating avoidance of infrastructllre by cnlving c:lribou.

'nese results suggest thai roads spaced too closelywill depress calving activity within the entire oil fieldcomplcx. In fact, relative occurrence of caribou in theheavily-developed western portion of the Kuparukpetroleum development area declined significantly from1979 through 1987, independent of total abullll:lnce (Fig.4.5) (Culleron et al. 1992b) .

U

20_All carlbou

• ,,' t=JCalvu

'"C• '",C

" .8;;;C

'"0.,U 0~

U. -0.4

<.8

·1.20, ,

Distance Interval (km)

Figure 4.3. Fraclional changes in mean densit1' or caribou from theCenlral Arctic herd bet,'1een pre·construction (1978-81) and post­conSlruction (1982-87) periods for 1-km·dislance intervals from theMilne Point road system in the Kuparuk petroleum development area,Alaska. (from Cameron el at. 1992b)

40 BIOLOGICAL SCIENCE REPORT USGSIBRD 2002-0001

90

80_ All Caribou

~ c:l Calvese70

0 60

I '0, 40.8~ 30•u.. 20

10

0- -79 eO " 82 83 84 85 " 87

Year

~4

• All caribou I.,....~O C Calves

0~

'":J 1.6

~l'.l 1.2

0ci 0.8 _,0.',.c .

'".. n'. ..~ 0.4 G. "'-. .. .. . . . .. . 'D'. .. '&-.-0'a-

0'079 51 82 83 84 " B6 87

Year

ro,.

"'"

"'.

".w

1979-81

1987-90

j.

1982·86

rlmlHl=h...

, .

'" "

0'.

,,,"

" . ".w

Figure 4.5. Decline In percentage abundance of canbou from theCentral Arctic herd west of the Mitne Point Road, Kuparuk petroleumdevelopment area, Alaska (Spearman's Rank, P < 0,02), and changesin tolar numbers of caribou observed north of the Spine Road (see Fig.4.3), 1979·1987, (from Cameron el al. 1992b)

Figure 4.4. Changes in mean relative distribution of caribou from theCentral Arctic herd in the Kuparuk petroleum dovelopment area,Alaska, during calving: 1979-1981, 1982-1986, and 1987-1990. Shownonly are those 10.4·km2·lransect segments in which the occurrence ofcaribou exceeded the area contribution to total coverage (0.9%).Gradations in line spacing depict multiples of observed use relative toavailability: \~ide =<3X; narrow = >3X-5X; solid =>5X. {from Smithand Cameron 1992}

An e.'\]lonelltial decline ill the occurrence of caribou asdensity of roads incre:lsed (Fig. ·1.6) (Ncllemann andC;ulleron 1(98) underscores the sensitivity Ilf the femalesdllring the calving period. The probable consequence isreduced access 10 preferred habitats (Bishop and Cameron1990, Ncllemann and Cameron j()96, 1993).

Incremental redi.~tribution anti local hahitatloss withintho.) Kupamk petroleulll development area may havetriggered changes on a regional scale. \VoIfe (2000)reported an inland shift in concentrated calving ,lctiYity

;t\vay from the I\lilne I\)int petroleum production unit(Fig. 4.7), apparelltly in reSponse to the increasing densityof infrastmcturc.

Ground observations within the Kuparuk petroleumdcvelopmcnt area in 1978-1990 provided additionalinsights on changing distribution and movements.Caribou incrcasingly avoided zones of intensive activity,especially during the calving period (Smith et al. 19(4),corroborating data from strip-transect surveys. Lowersuccess in crossing road/pipeline corridors by largeinsect-haras.~ed groups (Smith and Cameron 1985,CllfalOlo and !'vlurphy 1986, l\lurphy and Curatolo 1987,Murphy 1(88) may have contributed 10 a general shiftfrom Ihe central KupanJk petroleum development ;1rC:ltoperipheral areas with less slIfface developmcnt and humanactivity. Routes of summer moyement arc now primarilysoUlh of OliklOk Point and along Ihe Kuparuk Rivcrfloodplain (Smith e( ;II. Jl)y.lj.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES

Body Condition and Reproductive Performance

41

Figure 4.7. Shifts in concenlrated calving areas, Central Arcticcaribou held, Alaska, 1980-1995. (adapted from Wolfe 2000)

Thcsc rclationships link the nutritional consequcnccsuf changes in distrilllUioll tl) thc reproductivc success of,:ariboll of the Central Arctic herd. \\bt nf the

o >0.6-0.3 >0.3-0.6 >0.6-0.9Road Density (kmlkm2)

o

5

-" a~4.s~

.~ 3cwa~ 2a~·0ro()

1

An analysis of the summer distribution of radio­collared females in 1980-1993 (Cameron Cl al. 1995)suggests that caribou usc of the oil field region at PrudhoeBay has declined considerably from that noted during the1970s by Child (1973), Whilcet al. (1975), and Gavin(l97R). Caribou abundance within the main industrialcompli:x as well as cast-west movcmcnts through that areawere significantly lower than for other areas occupied bycaribou along the arctic coast (P = 0.001 and P < 0.001,respectively). Conservative calculations yielded anestim:ltcd 78% lkcreasc in usc by caribou and a 90%decrease in their lateralmovemcnts (Cameron ct al.l (95), all changes apparcIHly in response to illlensi vedevclopmcnt of the Prudhoe Bay to Kuparuk oil Heldregion over the past 3 decades. Occurrence of caribou thatusc the complex, however, is reportedly unrelated todistance froill infrastnJcture (Cronin et al. 19(8).

Figure 4.6. Relationship between mean (SE) density 01 caribou fromthe Central Arctic herd and road density within preferred ruggedterrain, Kuparuk petroleum development area, Naska, 1987-1992.Different lellers indicate a significant difference (P < a.OS). (fromNellemann arid Cameron 1998)

Reproductivc success of caribou is highly correlatedwith nutritlollal statuS. 'lllC probability of producing a calfvaries directly with body weight and/or f:lt content ofsexually-mature females during the prcvious alltullln(Cameron et al. 1993.2000; Camcron and Vcr I-locI' 1994;Gerhart ct <ll. 191)7). In cOlltra.q, calving date andperinatal survival arc more closely rclated!o maternalweight shortly after parturition (Camero-n ct al. 19(3)(Fig. ·1.8). The likelihood of conceiving is probablydetermined by botly condition at breeding. whereasparturition dale and calf survival relkct maternalcondition durin£! late gestation.

rI

III

42 BIOLOGICAL SCIENCE REPORT USGS/BRD 2002-0001

Reference ZoneA

0.30

q) 0,25c~ 0.20-~ 0.15 ..05" 0.100Z 005

:

···

···:..

P =0.2274..

B

0.00 -I-~-_---';'-_~_-'--:'-.-~

Hl80 1962 1984 1986 1966 1990 1992 1994 199tlYear

Treatment Zone

0,30y=23.186-0.0t16x

P '" 0.0001r'= 0.30

~ 0,15o>- 0.10oZ 0.05

Figure 4.8. logistic regres~ions (solid lines ale significant al P < 0.05)of parturition rale, incidence of early calving (i.e., on or before 7 June),'and perinatal (>2 days post partum) calf swvwal on autumn andsummer body weights of female caribou, Cenlral Arctic caribou herd,Alaska, 1987-1991. The empirical percenlages are shol'm al arbitrary10-kg intervals of body weight. Numbers in parentheses are samplesizes. The asterisk indicates indusion of one fomale weighing 57 kg.(from Cameron et al. 1993)

Sagavanirktok River, inlhe petroleum developmcnt zone,caribou had reduced access to preferred fnraging habitatsIlcar roads (Ndlelllalln and Cameron 1996) and shiftedtheir cOIlCelllratcd calving area into habitats with lowerplant biomass (I' < 0.001) (Wolfc 2(00). In contrast,foragc biolllass remained constant (I' = 0.23) withinconcentrated o::alving areas east of the SagavanirktokHiver where no (levc!opment was present (Wolfe 2000)(Fig. 4.9).

Repeated use of lower-quality I.:alvi ng habiwts mayrcduee forage illlake by females calving west of theSagavanirktok River. Likewise, impaired SUllllllermovelllelllS between insect rdid habitat and inl:H1dfeeding areas could depress ellergy halance (Smith 1996)and, hence, rates of weight~gain.

Indeed, sl:veral data sets suggest rl:dllccd nutritionalstal\lS and fecundity of radio-collared females exposed tooil deve lopmel1t west of the Sagavanirktok River.Estimates of July :llId October hody weights, over­SllJlllller weight-gain, thc incidence of 2 successive-yearpregnancies, and jk'rinatal calf slll"\'lv:l1 all tended 10 be

---z . 1 ;~.~om ••

HMO 1982 1984 1986 1968 1990 1992 1994 1996Year

Figure 4.9. Changes in median Normalized Difference VegetationIndex (NOVI} on 21 JUrle for concentraled C<lIYing areas 01 the CentralArctic caribou herd in the study reference zone (relativelyundeveloped) and treatment zone (developed} easl and west of theSagavanirktok River, Alaska, respectively, 1985-1995. (from Wolfe2000)

lower for fcmales to the we,,1 than for those underdisturbance-free conditions to the cast, aHhoughindividual differences were not significant at the 95%confidence level (Cameron et al. 1992a).

In a more rccent analysis of data for 1988-1994,howcver, mean parturition nlte of females I.:alving west ofthe SagavanirklOk River was less than that of femalescalving 'cast of Ille Sagavanirkwk River, 6V.1o vs. 83%,respectively (I' = 0.003, Table 4.1) (C':Hneron 1995).Corresponding frequencies of repmdudive pauses(Cameron 1994, Cameron and Vcr Hoef 1994) wercsignificantly higher (I' < o.m, I-test, ratio method) in theweSt (36<;~" 26 of 73 observations) compared with the cast(19'.";', 12 of 64 observations), IJr approximately one pauseevery 3 and 5 years, respectively (Cameron 1995).

The key constraint on repfllduction is lactation, whichexal.:!s a substantial I.:ost Oil sUll\mer weight-gain, in turninnuenclng the probability of conceiving that autumn.During 1988-190 I, weights of alllact:lllng Central Ardll.:herd femaks sampled :lveragL'd 9 kg less than

ARCfIe REFUGE COAS'I:<\L PLAIN TERRESTRIAL WILDLIFE RESEARCH SU1'.'[1'.IARIES 43

·-......PottunbO<lR~!•• O.ell

Autumn Body WoJOht (kll)

.i· .. ·......

, ...

1996-2000 (Fig. 4.2) Sllgg~sls the prevalence of forageand inscct conditions that cnhanced growlh and f:lll~ning

dcspite the d~Jllands of milk production and presence ofinduslrial aClivity. With the opening of the Badamipelroleum development area cast of the SagananirkrokRiver in 1996, howevcr, the undisturbed status of lhalarca was compromised. rendering funher comparisonsqll~slionable.

Clearly, anthropogenic impacts on caribou mllst beidentified and :Iss~ssed within the framework of a variablenatur:ll environment. Favorable foraging and insectconditions would allenuale lhe conseqllenc~s ofdisturbance-induccd changes in quality of occupiedhabilats. Conversely, adv~rsc conditions would exacerbatethose same lypes of consequences. Unless analys~s archased on multi-year observations of marked individualsand incorporate comparative data on all undisturbedcontrol or reference group. conclusions will be equivocalal best. For example, absent a valid baseline, net growthof the Central Arctic herd (Fig. 4.2) is no b~H~r evidence{)f compatibility wilh developmenl lhan a net declinewould be evidence of a conflic\.

111e crucial consideralion for the fUlure of lhe CcntralArctic herd and other arclic caribou herds is whetherchanges in distriblllion associated with surfacede\'<.~loplllent, by depressing repmdtKtion \)r survival, willeither reI ani an incrcase in herd si1.c or accelerate adecrL'asc.

Our data, in fact, indicate that productivity can andwi II decline if the clllllulalive loss of preferred habil:lt,when superimposed on natural forces, is .~ufficient tocompromise Ilutrition.

FIgure 4.11. Dislributions of observed autumn (Oelober) body weightsfor laclating and nonlactating female caribou from the Central Arcticherd. The associated parturition rales are inlegrated estimates derivedfrom lhe logistic model (Fig. 4.8). (from Cameron and While 1992)

Overview

% Parturient (n)

Year West East

1988 72.7 (ll) 100.0 (8)

1989 53.8 (13) 77.8 (9)

1990 83.3 (12) 100.0 (7)

1991 45.5 (11) 75.0 (12)

1992 72.7 (11) 75.0 (12)

1993 55.6 (9) 62.5 (8)

1994 66.7 (6) 87.5 (8)

Table 4.1. Parturition status of 43 radio-collared female caribou8,Centra)Arctic herd, west and east of the Sagavanirktok Riverb, Alaska, 1988­1994. West includes the Prudhoe Bay and Kuparuk oil fields; easl wasgenerally free of disturbance during thai time. (data from Cameron 1995)

a All sexually malura.

b rndlviduallocaUons conslstonlly west or cast lor 2-7 years duringtho calving period.

Host, paired comparisons, P=O.003.

llonlacl:lting fcmales (Fig. 4.10). This resulted in aprojectcd 28% lower parturition ratc for the laclatingfemalcs (Fig. 4.11) (Carm:ron and Whitc 1992).

Lower parturition rates of fell1<1les wc-st of theSagavanirktok River during 1988-94 (Table 4.1) mayreflect a failure \0 compensale for the metabolic burden ofmilk production (i.e., through increased forage intake orreduced energy cXJlcnditure). Hence, those fernaks of theCc-ntral Arctic herd that Ilsed thc dcvelopmcllt zone werein consistently poorer condition in autumn, expericnt:edllIOTe frequcTlt reproductive pauses, and produced fewercalves (Fig. 4.2).

Yet the degree to which laclation constrains weight­gain docs vary. An increase innel calf production during

Moan parturlUon rato% 64.3'·± 5.0 82.S";!; 5.3

Figure 4.10. Mean (SE) body weights of lactating and nonlactatingfemale caribou from the Central Arctic herd. Alaska, in summer (July)and autumn (October). (from Cameron and White 1992) 'Significanl atP< 0,001.

'00 r0 lloo1oct>6ngIG Llt!.,U.1(I r-:l- ... " ... '.

00 • _ !lk~

"" .

".-

" ~J- ""~ ••• 00'

~

I •••

'" IliB ~,,'---- .__ ., ----Summar Autumn

rI

II

44 BIOLOGICAL SCIENCE REPORT USGSIBRD 2002-0001

References

Ad:lfllC7_Cwski, J. Z., C. C. Gates, R. J. Hudson, and 1\-1. A, Price,1987. Seasonal changes in body composition of maturefelllak caribou and calves (Rdllgifcr (ilf<1JldIiS

grow/andicus) on an :lrctic island with limiteJ winterresollrces. Canadian Journal of Zoology 65: 1149-1157.

Bishop, S. c., and R. D. Cameron. 1990, Habitat usc by post­parturient femak caribou of the Central Arctic herd. Paperpresented at the annual metting of'I11c Wildlife Society.:\l;lska Ch;lJlter. Juneau, Alask:!, USA, ·1·6 April 1990. Bookof abstracts:9.

Cameron, R. D. 1994. Reproductive p:luses bv fern:llc caribou,J(lunlal of 1>1:lll1lJlalogy 75: 10-1 3. •

__. 1995. Can petroleum development depress theproductivity of arctic e:lribou? Paper presented at theSecond International Arctic Ungulate Conference,Fairbanks, Alaska, USA, 1:1-17 August 1995. Bllok ofab.>lracts:36.

_.__. S. G. Fancy, and W. T. Smith. 1992a. ReproductiveperfoTllm}ce of caribou in relation to habitat a\':libbilit\' andquality. Pages 67-78 in T. R. MeC:lbe, n, Griffith, N. E~Walsh, and D. D. Young, editors. Rew:lrch on the poteillialeffects of pClrokum development on wildlife and theirh:lbitat, Arctic National Wildlifc Refuge interim report,J988-1990. Alask:l Fish and Wildlife Research Center andArctic National Wildlife Refuge, Fairbanks, Alaska, USA.

__, E, A, Lenart, D, J. Reed, K. R. Whitten, and W. T.Smith. 1995. AbunJance and movements of caribou in tbeoil Held complex ncar Prudhoe Bay, Alasb. Rangifer15:3·7.

__, D. J. Reed, J. R. D:m, and W. T. Smith. 1992/1.Redistribution of calving caribou in response 10 oill1eldtlevclopmelll 011 the Arctic Slope of Alaska. Arctic 45:338­342.

__, O. E, Russell, K. L. Gerhart, R. G. White, and J. M. VerHod. 2(X)O. A model for predicting the parturition status ofarctic caribou. Ran£!ifer, Speeial Issue 12: 1-3.

__, W. T. Smith. S. G. F:lney. K. I.. Gerhart, and R. G.White. 1<)93. Calving success of female c:lribou in relationto body weight. C:lIladian Journal of Zonlogy 71:-180-486.

__, :H1d 1. !"II. Vcr Boef. 1994. Predicting parturition ratc orcarib{)u from autumn hody mass. Journal of WildlifeManagement 58:67·\·679.

__, and R, G, White, 1<)92. lmp,m:lnce of summer weightgain to the reproductive sUCCeSS of caribou in Arctic Alaska.Paper presented attlie Fifth Australasian Wildlifel\!:m:lgelllellt Society Conference, 2·4 Decemher 1992Brishane. Australia.

__' :Illd K, R. Whillen. 19S0. Intluence of the tr:un;- ..\Ia,bpipeline corridor on the local dbtrihution of c:lrilH)u. I':lges·175·-18·\ in E. Reimers, E. G:l:lre. :ll1d S. Sl;jenneberg,editors. l'rocl'edillgs of Ihe Second Inkrnation:ll R<,ind,'l'T1Caribou Sylllposium, Rows. Norway. Direkwratct for viiiog ferskl'annsfisk, Twndheim. Norway,

__, __, and W. T. Smith, and D. D. Roby. 1979. Cariboudistribution and group composition associ:lted withconstmetion of the trans-Alaska pipeline. Canadian field·Naturalist 93: 155·162,

Child, K. N. 1973. The reactions of barren·ground c:lribou(Rilllgifer lari1lldus gnlllli) to simulatcd pipeline andpipeline crossing stmetures at Prudhoe Bay, Alaska.Complelion Report, AI:lska Coopef:ltivc Wildlife ResearchUnil, University of Alaska, fairbanks. Alasb, USA.

Clough, N. K., 1'. C. Patton, and A. C. Christiansen, editors.1987. Arctic National Wildlife Refuge, Al:lsb, coastal pl(linresource assessment - report and recolllmendation to thcCongress of the United States and final legislativeenvironmelllal illlpact statement. U.S. Fish and WildlifeService, U.s. Geological Suryey, and Bureau of LmdManagement, \Va..shington D. C.

Cronin, M. A., S. C. Amstmp, G. M. Dumer, L. E. Noel, T. L.McDonald, and W. B. Ballard. 1998. Caribou distributionduring the post-c:l);'ing peri oJ in relation to infraSlmcture inlhe Pmdhoe Bay oil field,/\lasb. Arctle 51:85-93.

Cur:llUlo, 1. A., and S. M. Murphy. 1986. 111e effects ofpipelines, ro;lds, and traffic on llW\'elllelllS of caribou,Rangifa lamlldlls, Canadian field-Naturalist 100:218-224.

Dall, J. R., and R. D. Cameron. 1986, Effects ofa road systemon caribou distribution during calving. Rangifer, SpecialIssue 1:95-101.

Dauphine, T. C" Jr. 1976. Biology of the K:uninuriakpopulation of barren-ground caribou. Part 4: growlh,reproduction, and energy rc.~crves. Canadian WildlifeService Report Series 38.

Elison, G. W., A. G. Rappoport, and G. M. Reid, ediwrs. 1986.Report of the caribou impact analysis workshop, ArcticNational Wildlife Refuge. 19·20 Nov 1985. U.S. Fish and\Vildlife Seryiee, Fairbanks, Alaska. USA.

Eloranta, E" :md 10.1. Niemincn. 1986. Calving of theexperimental reindeer herd in Kaam:men during 1970-85.Rangifer, Special Issue 1: 115- J21.

Gavin, A. ca. 1978. Caribou migrations and p:llleTlls, PmdhocBay region, Alaska's North Slope (1969'1977).Unpuhlished report, ARCO Alaska.lnc., Anchoragc, Alaska,USA.

Gerhart, K. 1.., D. E. Russel1. D. V:1I1 de Wetering, R, G. White,and R. D. Cameron. 1997. Pregnancy of adult caribou(RallgiJ~'r IllTamill.l"): evidence for lactational infertility.Joumal of Zoology 2·12: 17 -30.

Haukioja. E., :lml R. Salo\'aar:l. 1978. SUllllller weight ofreindeer (Rililgifa lilrillliJlIs) call'"s :lnd its illlport:mce fortheir fUlllrc survival. Report Ke\"l) SUbarCIl<; RescarchSlation 1·1: 1-4.

Lellyik, D" O. Grimef:iel!. ami J. Tamnes. 19HH. Selectionstf:ltegy in domestic reindeer: 5. Pregnane" in domcsticrcimJccr ill Tnllldcl;lg C'oullty. Norway. N(;r,kLandhruksfmsk. 2:151-j61.

M1I1Vhy. S. M. 1988, Caribou behavior and n)()WnJ<,nts in theKuparuk oil field: impli\.~alions for energetic :lJld imp:lctanalyses. Proceedings of thc Third North ..\mcrican Carib<1u\\'orkshop< Alaska Departmelll of Fish and Gmnc, Juneau,Al:\ska, US..\. Wildlik Technic:ll Bulletin 8:19(}·21O.

i ARCrIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU/I.·H.fARIES 45

__, :llld J. A: Curatolo. 1987. Activity budgets :mdmo\'<:mcnt r:lles of caribou encountering pipelines, roads,and traffic in Northern Alnska, C:llladian Journal of Zoology65:2483-2490.

Ncllcmann, C. :Ind R. D. CUncf\)ll. 1996. Effects of petroleumdevelopment on terrain preferences of calving caribou.Arctic 49:23-28.

__• and __" 1998. Cumulative impacts of:ll1 el'olvingoil-field complex on the distribution of calving caribou.Canadian Journal of Zoology 76: 1425-1430.

Reimers, E. 1983, Reproduction in wild reindeer in Norway.Canadian Jounwl of Zoology 61 :2\ [-217.

Rognmo, A., K. A. l\farkusscn, E.l:1cobscn, H. I. Gr:l\', and A.S. mix, [983. Effects of improved nutrition in pregl1~lJl

reindeer on milk qu:llity. calf birth wcight. growth. andmort~lity, R~ngifcr 3: 10-18.

Skogl~l1tl, T. 1984. 111C cffceL~ of food and maternal condilionson fctal growth and 5i/.c in wild rcindeer. Rangifcr 4:39·46.

Smith, M. D. 1996. Distribution, abundance, and quality offorage within the summcr rangc of the Ccntr~l Arcticcaribou herd. TI1csis, University of Alaska. Fairbanks,Alaska, USA.

Smith, W. T., and R. D. Camcron. 1983. Rcsponses of caribouto industrial developmcnt on Alaska's Arctic Slope. ActaZoologica Fennica 175:43-·15.

__, and __. 1985. Reactions of l~rgc Ernul'S of caribouto a pipeline corridor on the nrctie coastal plain of Alaska.Arctic 38:53-57.

__, and __. 1992. Caribou responses to developmentinfr~structurcs :ll1d mitigation measures implemented in theCentral Arctic Rcgion. Pages 79-86 ill T. R. !'.fcCabc, B.Griffith, N. E. Walsh, and D. D. Young. cditors. Hcse:lrch Oil

the potential effccts of petroleum developmcnt nn wildlife:ll1d their habitat. Arctic National Wildlife Rcfuge interimreport, 1988·1990. Alaska Fi>h and Wildlife RcsearchCenter and Arclie Nation~l Wildlife Refuge, F:lirbanks.Alaska, USA.

__, __' and D. J. Rced. 1994. DistribUlion andmovements of caribou in relation to roads and pipelines,KUp:Ullk Development Area, 1978-90. Alaska Department ofFish and Game. Juneau, Alaska, USA. Wildlik TechnicalBulletin 12.

'nlOmas, D. C. 19H2. TIle relationship between fertility :md fatre~crves of Peary caribou. Canadian Journal of Zoology60:597·602.

__, and H. P. L. Kili;lan. 1991. Firc·earibou relationships:II. Fecundity and physical condition in the Beverly herd.Unpublished, report, Can:ldian Wildlife Service, Edmonton,Alherta, Canada.

White, R. G. !983. F()ragin~~ pallerns and their lllilitiplier effectson produnivity of ntlrthcm ungulatcs. Oikos ·10:377·3/;4.

__, n. R. Tholllson, T. Skoglaml, S. J. Person, D. E, Russell,D. F. Holleman, :lIHI J. R. Luick. 1975. Ecology of l~aribolJ

:ll Pnldhoc Bay, Alasb. Pages 151-20 I ill J. BlOwn. editor.Ecologic:l! investigations of the tundra Iliume in thePrudhoe Bay rl,gion, :\la~ka. lliologicaj-J)al'ers, Universityof Alaska, Special Report 2.

Whitten, K. R. and R. D. Call1cron. 1983a. Movements ofcollared c<lribou, Rmzgijer /(lrimdus, in relation 10 petroleumde\'clopmcnt on the Arctic Slope of Al:lsk~. Canadi~n Fie1d­Natur~list 97: 1·13-146.

__, and __. 1983b. PopUlation dynamics of the CentralArctic herd, 1975-1981. Acta Zoologica Fenniea 175: 159­161.

__• and __. 1985. Distribution of caribou e~lving inrelation [() the Prudhoe B~y oil field. Pagcs 35·39 ill A.Martell and D. Russell, editors. Caribou ~nd human ~ctivity:

proceedings of the First North Amcrican Caribou Workshop,1983, Whitchorse, Yukon Territory. Canadi~n WildlifeService Speci~l Publications, Ottawa, Ontario, Canada.

Wolfe, S. A. 2000. Habitat selection by calving caribou of theCentral.·\rctic herd, 1980·95. Thesis, Univcrsity of Alaska,F~irbanks, Alaska, USA.

46 BIOLOGICAL SCIENCE REPORT USGS/BRD 2002-0001

Section 5: Forage Quantity and Quality

Janet C. JOI:q •.'II.\OI/, Mark S. Udt'l'it:., alld Newey A. Felix

The Porcupine caribou herd has tr:lditionally used thecoastal plain of th~ Arctic National Wildlife Refuge.Alasb. for calving. Availability of nutritious forage hasbeen hypothesizcd as one of lhe reasons the Porcupinecaribou herd migrates hundreds of kilometers 10 reach thecoastal plain for calving (Kuropat and Bryant 1980.Russell et al. 1993).

Forage quantity and qual it)' ;}nd the chronology ofsnowmdt (which determines availability and phenologicalstages of forage) have been suggested as important habitatattributes lhat lead calving caribou to select one area over,wother (Lent 1980, White and Trudell 1980, Eastland etal. 1989). A majorqu~stion when considering the impactof petroleum devdopment is whether potentialdisplacement of the caribou from the 1002 Area toalternate c:llving habitat will limit access to high quantit),and quality forage.

Our study had the following objectives: I) quantifysnowmdt patterns by area~ 2) quantify relationshipsamong phenology. biomass, and nutrient content ofprincipal forage species by vegetation lype; and 3)detemline if traditional concentrated calving areas differfrom adjacent are:ls with lower calving densities in termsof vegetation characteristics.

We investigated caribou forage in 2 areas: anhistorically tr:lditional calving area entirely within the1002 Area and :In :ldjacent displacement area entirelyoutside of the 1002 Area (Fig. 5.1).

111e traditional calving area was defined during the1002 Area baseline biological studies as the area ofintensive calving use during [0 of the 14 years studiedfrom 1972 to 1985. Importance of this area was upheld bydata on calving locations from later years.

Availabil ity of potential displ:lcement areas is limitedhere becausc the eoast:l1 plain narrows as lhe rugged

'-, ~ " "

Figure 5.1. Map of c.3ribou forage sludy area. Porcupine c.3ribou herd,on the coastal plain ollhe Arctic National Wildlife Refuge, Alaska.

Bmoks Range mountains approach the Beaufort Sea. Thedisplacement :lre,L cho.'ien for comparison was located tothe south and east of the 1002 Area in the only [Jan ofAlaska's North Slope thaI is a designated WildemessArea.

The displacelllel1l area had topography similar 10 lhelraditional area: a mix of rolling foothills and Ct);lsta[plains. It lay along lhe spring migration route typiCilllytraversed by the Porcupine caribou herd. Female caribouhave calved in this area. especially during years whensnow melled late (Fig. 3.18).

We gathered data at both study areas during thecaribou calving period in early June of 1990 and 1991.Fifly-five 30m x 30m study sites in 1990. and 45 in 1991,were loc:lted at the intersections of grids positionedr:lndomly ovcr the entire study area. We sampled during 3periods in 1990 (311\Iay-) June. 8-12June, :ll1d 19-22June), and 2 periods in 1991 (4-6 June and 9- J3 JUlie).

Data were collected on 4 prevalent plant speciesidentified in the literature as impL)rtant forage for caribouon Alaska's North Slope: ErioJl/1orum vllgil/(l/um (lUssockcottongrass). E. (ll/glIstiloliulII (tall eOllongrass). Cart'.\·at/limitis (aqu:ltic sedgcl, and SlIli.\' pfmrijolia ssp. Jlllfdmi(diamond-leaf willow) (Thompson :ll1d !'.1cCourl 1981,Russell et al. 1993).

At each study site. for:lge quantity data were collectedin 14 I-Ill: quadr:lts along 2 randomly-located transects.Phenology data were collected at the same 14 quadr:lts,plus 20 3-m: quadrats located on 2 additional randomtransects, Tussock cOllollgrass inl10rescences anddiamond-leaf willow \ca\'es were colleeled on transectsfor analyses of nLJlriel11 and fiber content at a r:lndolllsubsamp\c of sites during 1990 only.

Wc compared the tr:lditional and displacement calvingareas by measuring thc following characteristics:distributions of vegctiltion types, snowcover. plantbiomass, nUlri~nt and fiber eOlllen\, ;md phenology.

Non"p:lramctric analysis of variancc using a repeatedmeasures design was used to test for differences betweenthe calving are:1S and between sampling periods (time).Analyses were conductcd individually for each parameterin each year.

Inter,Lctions between area and tillle were tested with1\-1anll-Whitney tests (Conover 1980) of area differenccsfor each time contrast ill a full orthogonal set. Ifinteractions were insignificant (/' > 0,(J5), area differenceswere tested wilh Ivlann- Whitney tests based on the Illeill\valuc for all s:llnpling periods. The differences betweenthe first :llld last sampling periods were tested with thesign test using data from both areas. If any inlaaetionswere significant, tests for differences between areas wereconducted separately for eilch sampling period, and signtests were conducted sep:lr:ltely for e:lch arei:!.

ARCfIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 47

Forage Comparisons Within and Outside the1002 Area

Spring snowmelt was very e:lrly in 1990 and wasIlearly complete before the calving period. Snowmelt in1991 was slightly earlier than normal. In 1991, we foundthe traditional calving :!rea had more lIfca with partialsnowcover remaining during the calving period than thedisplacemelll area (40<;';' vs. 33%, P = 0.02). Some plalllswere in earlier phenological stages in thc traditional area(Table 5.1).

TIle quantity of new green forage was low throughoutboth study areas during the calving period in both yearseven though snow melted earlier than normal (Table 5.2).Only tussock cOllongrass flowers appeared 10 be readilyavailable for forage during the peak of calving. 111e 2other sedges had very little new growth. The willow,which is the major forage species latcr in the summer,leafed out only at the end of the calving period. Thetundra appeared brown and no other plant species wereproducing abundant new growth during the calvingperiod.

Four forage species were quantified Cfable 5.3).Tussock cottongrass !lowers had much higher biomass inthe traditional calving area than in the displacement areain both years, although the difference wa~ statisticallysignificant only in 1991. Tussock coltongrass was

uncommon at the displacement area sites. The 2 othersedges had very low cover and no significant differences.Diamond-leaf willow did not leaf out during the studyperiod in 1991, but in 1990 willow leaves had gre(lterbiomass in the traditional calving an~a when they leafedOUI.

Tussock cotlongrass flowers and diamond-leaf willowleavcs were tested for forage quality throughout the slUdyperiod in 1990 (Table 5.4). Higher nutrient concentrationsincrease forage quality, while higher fiber and ligninconcelllrations dccrease digestibility. Both plant spcciestendcd to have greatcr forage <juality ill earlierphenological stages than in later stages. Tussockcottongr:lss flowers had greatcr forage quality in thetradition:ll calving area than in the displaccmellt area.

Distributions of vegetation types were distinctlydifferent betwcen the 2 areas (fable 5.5). '111e traditionalcalving area had greater cover of 2 vegetation typesimponant for caribou forage: tussock tundra and moistsedge-willow tundra. The displacement area had greatercover of early succession vcgetation types such as Dryasrivcr terr<lces and barrcn or panially vegetated ground dueto the greater extent of !loodplains in that area. Tussocktundra is a late-succession vegetation type and is nearlyabsent from lloodplain terrain (Jorgenson et al. 1994).

The displacement area also included the highestelevation foothills where development of tussocks is

Table 5.1. Median phenological slagesaor major forage species in the Porcupine caribou herd's lraditiorlal caribou calving area (C) and potentialdisplacement area (D) on lhe coaslal plain of lhe Arclic Nalional Wildlife Refuge, Alaska.

1990 31 May- 03 Juno 08 - 12 June 19-22Juno

C D C 0 C 0 Amab Tlmoc

Tussock coltongrass 3 3" d 3 3"d 3.. P.:O.Ol4

Tall collongmss 2 2 2 2 P"O.60 "Aquatic sedgo 2 2 2 .0=0.44

Dlamondloaf willow .. .. P"O.Ol2 2 2 3

1991 04 -06 Juno 09-13June (no sampling)•r c 0 c 0 Aroab Tlmac

Tussock collongrass 2 2 3 3 P=O.85

I Tall collongrass .i P=O.04 "• 2i',

Aquatic sC!dge 2 2 P::O,83 "'Dloffiondtoaf willow 0 O"d 0 P.:O.Ol

• P< 0.05, •• P.: 0.01, ns _ no significant dilforence., Phenological stages: Tussock coltongrass, 1 := boot stage, 2 := early flower, 3 := lullllower, 4 := s(){)d; Tall cottongrass andAquatic sedge, 1 := vegetatlvo .: Scm, 2 := vegetative;> SCm, 3 := early 1Iower, 4 := full nower; Diilmondleaf willow, 0 == no nowgrO\~1h. 1 == bud swollon, 2 := leaf unfolding, 3 := full lea"

b SlgniHcallCo lovel for difference between oreas averaged over all periods., Significance level for chango boN/eon first and last sampling period.d Phenological stage signllicanUy i'luvanced, allhough median valuos wore lhe same.

48 BIOLOGICAL SCIENCE REPORT USGSIBRD 2002-000J

Tabla 5.2, Medianabiomass (g/m2)and percent cover of 4majorcaribou forage species in the 5 mosl common vegetation types on thecoastal plain of the Arclic Nalional Wildiife Refuge. Alaska. during thePorcupine caribou herd's calving penod. June 1990.

glm~ 0.00" 0.881le O.OOJJl O.83c 3.2Sc

a Values are medians for all siles. with site vatues obtained asmoans for 3 times. Medians with the sarno supem::ript werenol slgnificanUy differont (P.:: 0.05).

b Vegetalion types; WG ::: weI gramlnoid tundra. MS ::: moistsedge-wiflow tundra. MSD ::: moist sedge-Dryas tundra. IT:::tussock tundra. ST::: low shrub tundra.

Tall COl1onqrass

% cover 1.2A1l 1.4" O.Se O,4e O.Sue

Aquatic slX1q!)

% cover 0.9' 0.,' O.OB 0.0" O.OB

DjamQndloilf willow

poorer than inlhc lower foothills of t}lC traditional calvingarea. Glacial lobes covered about olle·fiflh of the foothillsin the displacement area during Ihe most recent glaciationthat ended -10.000 years ago. lllese recelllly deglacialedareas have very little tussock lundra (Jorgenson 1984).

Different distributions of veget3tiontypes may explainIllost of the differcnces found in fornge quantity andquality bctween the traditional and displacement areas.Tussock cotlollgrass nowers had greuter biomass andforage quality in the tussock tundra type compared withother ,'cget3tioll types (Tables 5.2 and 5.6). The greaterbiomass of flowers in the traditional a.rea mainly resultedfrolllthe greater eXICnt of tussock tundra. No consistentdifferences in flower quantity or quality between tussocktundra in the traditional calving an~a ;llld tussock tundra inthe displacement area were observed.

The location of the Porcupine caribou herd traditionalcalvin!! area is greatly influenced by veget:llion typedistributions and snowmelt patlerns across the ArcticRefuge coasl:ll plain. lllese factors appear to detenninethe quantity and quality of forage availablc to Ule caribouduring calving.

Because of the low amount of for:lge available duringthe calving period. the differences in vegetation

STITMSD

.:: O.OlAlJ 0.04e 0.031lC

VegataUon TYPesb

MSWGSpecies

Table 5.3. Median densily (no/m2). biomass (gfm2), and percerll cover of major forage species irlthe Porcupine caribou herd's lradiliorlai cariboucalving area (C) and polpnlial displacement area (0) on Ihe coastal plain of the Arctic NatlQnal Wildiife Refuge. Alaska.

199031 May- 03 June 08-12June 19 - 22 Juno

C D C D C D Area!! Tlmeb

Tussock collongrass flowers

nolm2 0.6 0 0.5 0.2 0.2 0 P=0.87 "'o'm2 0.02 0 0.01 <: 0.01 0.D1 0 P=1.00 n'

Tall collongrass

% cover 0.' 0.4 0.6 0.3 0.6 0.2 P:::0.93 '"Aquatic sedge

% cover 0 0 0 0 0 0 P=Q.72.,

Dlamondlear willoW leaves

o'm2 0 0 0 0 2.0 0.3 P=O.10

199104 ~ 06 June 09-13June (no sampling)

C D C D C D Areaa Timeb

Tussock collongfllss flowers

no/m2 2.7 0.0'· 3.5 0.0·· P<:0.01 n'o'm2 0.09 0.00·' 0.11 0,00·· P<:O.Ol n'

Tall cottongrass

% COver 0.2 0.3 0.3 0 P=O.68 n'Aquatic sedge

% cover 0 0 0 0 P=Q.oa .,• P <: 0.05, •• P <: 0.01. liS no significant difference.4 Slgnillcance levol fQr difference bel\veen areas avoroged over 311 periods.b Significance Jevel for change between first and lasl time periods. If the significance levels diflorod betwoen areas. both are

shown (C/D}., last sampling period with significantly hlghar cQvor, although median valuos were tha sarna.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU1'.lMARJES 49

Table 5.4. Median nlltrient and fiber concentrations {percent of dry weight) of 2 major forage species in different phenological stages in thePorcupine caribou herd's traditional caribou calving area (e) and potenlia! displacement area (D) on the coastal plain of-the Arctic NationalWildlife Refuge, Alaska.

Early Flower Full Rower Sow

Tussock coltongrass C 0 C 0 C 0 Areaa Tlmab

Nitrogen 2.8 2.6..

2.2.. P=O.Ol "'2.5 2.2 2.4

Phosphorus 0.49 0,45 0,46 0.38.

0.36. P:0.03 'Ins0.38

Calcium 0.18 0.16 0.13 0.12 0.14 0.11 P=0.25 "'Neutral detergent fiber 55.8 58.0 58.2 62.7 66.8 65.9 P:0.07

AcId detorgent fiber 16.5 18.4 19.0 ..22.8

.. P:O.Ol20.4 23.2

Ugnln 2.3 2.6 2.5 2.3 1.9 1.9 P==O,43 "'n (II of slles) 6 7 8 15 8 14

leaf Unfolding Full leaf

Dlamond/eaf willow C 0 C 0

Nitrogen 3.7 3.4 2.2 2.2 P=O.27

Phosphorus 0.50 0.47 0.18 0.17 P=0.79

Calcium 0.51 0.49 0.62 0.59 P=Q.34

Neutral detergent tiber 22.2 22.4 25.0 25.6 P=0.71

Acid detorgont fiber 16.3 15.6 17.0 17.7 p..O.71

lignin 10.3 9.4 8.8 9.4 p,.O,43

n (II of siles) 8 8 8 8

, P< 0.05, .. P< 0.01, ns _ no significant dilference.0 Significance lavel for difference between areas averagod over all periods.b SIgnificance level for chango between first ond last lime periods. If tho slgnifk:anco levels differed between areas, both are

shown (C/O). A1ltosts for tussock cottongrass Inflorescences wara of lull flower and soed stages only.

Table 5.5. Dislribulion ofvegelalion types (percent of area) in the cariboll habitlt study area (Fig. 5.1) based on an independent sample of756systematically-located vegelation plots on the coastal plain of the Arctic National \Vildlire Refuge, Alaska.

Vegetation Type Entire Calving Habitat Traditional Displacement CalvingCoastal Plain StUdy Area Calving Area Area

Tussock Tundra 22 30 39 21

Moist Sedge·Willow Tundra 30 25 31 19

lolV Shrub Tundra 7 11 8 13

Moist Sedge·Dryas Tundra 12 9 7 12

Wet Graminold Tundra 13 8 8 7

Dryas River Terrace 3 7 13

Riparian Shmblands 2 3 3 3

Barren 2 3 3 4

P<lr1ially Vegetated 2 2 <1 5

Water <1 2 <' 4

50 BIOLOGICAL SCiENCE REPORT USGSn3RD 2002-0001

Table 5,6, Median nutrient and fiber concentralions (percent of dry weigh!) of tussock collongrass inflorescences in different phenological stagescompared oolvieen tussock lundra and o!her vegetation types on !he coastal plain of the Arctic National \'Vildlife Refuge, Alaska.

Full Flower Sood

n" Other n Olher Stage

Nilrogen 2.4..

2.4 Po:: 0.012.0 2.1

Phosphorus 0.43..

0.39..

0.34 0.35 P 0:: 0.01

Neutral detergent liber 60.0 63.6 66.2 66.7 P '" 0.25

AcId datergent fiber 19.7 21.8.

22.8.

P""0.0423.6

lignin 2.6 1.1 1.9 1.9 P", 0.03

nlilolsiles} 17 6 16 6• P 0:: 0.05, ··P<.O.Ol

a IT '" tussock tundra; Other'" all other vegetallon typos that had tussock collongrass inflorescences lrlcluding moist sedge-willowtundra, moist sfrdge-Dl)'as tundra, and shrub-dominated tundra.

characteristics found between the 2 areaS studied arcprobably biologically impOrtillll and could affect cnlving

success of the herd if pctroleum development causes thc

displacement of c,llving caribou Out of the 1002 Area into

adjacent arcas with lower forage valuc,

References

Conover, W. J. 1980. I'ractic:ll nonpararnctrie ~talislics. Secundcdition. John Wiley :md Sons, New York, Ncw York, USA.

Easlland, W. G.. R. T. Bowyer. and S. G. Fancy. 1989. EffeclS ofsnow cover on selection of calving .~ile~ hy caribo\!. Journalof Mammalogy 70:112'1-828.

Jorgcnson, J. c., P. E. Joria, 1'. R. l\lcCahe. ll. R. Reill., 1\1. K.Raynolds, M. Enwrs,l\1. A. Wilms. 1994. Users guide forthe land·cover nwp of lhe eO:lstal plain of lhe ArcticNational Wildlife RctilgC. u.s. Fish and Wildlife Service,Fairbanks, Alask:!, USA.

Jor.gellson, M. T. 198·1. 'nle response of I'egelalioillo !:mdscapeevulution on gl:;cial till ne;lr Tonlik L1ke, Alaska. Pages13·1-141 ill V. J. Lallau and C. L. Kcrr, editors.lnyentoryingforcst and olher veg'~lati()n of the high .-.1tilu(/c regions:proceedings of all irlleOl:uionalsymposiurn, Society ofAmeric:1ll Foresters Hegion.-.I Technic.-.! Conference,F:rirbanks, ..\laska, USA.

Kuropat, P., and J. P. Bryant. 1980. Foraging heh:wior of COI\'caribou on lhe Ulukok calving grounds in northwcsternA1ask:!.l'agc.~64-70 ill E. Heimers, E. Gaare, lmd S.Skjennehcrg, ediwrs. Proccedings of the SecondInlernational Heimkcr/Carihou Symposium, Hows, Norway,JI}79. lJirekhJralCl for vill og ferskyannfisk, Trondhcirn,Norway.

L<.'nl, 1'. C. 19S0. Syrwjltic ,no\I'lIl<.'h p:1l1crns in arctic ,\lasb inrelation to cllribnu habi1111 use.Pag,·s 71-77 ill E. Reimers,E. Gllllre, and S. Skjenncberg, dilors. Proceedings of lheSecond 1I1tcn!atiO!lal Reindeer/Caribou Symposium. HOf()s,Norway. 1979. DirektoTaICI [(ir vill og ferskvannfisk,TroJl~heirn,Norway

Russell, D. E., A.I\1. Martell, and W. A. Nixon. 1993. The rangeecology of the Porcupine caribou herd in Canada. Rangifcr,Sp.;:cial bsuc 8.

Thompso!J, D. C., and K. II. McCourl. 1981. Seasonal diets oflhe Porcupine caribou herd. American lIfidland Naluralist105:70·76.

While, R. G., :md J. Trudell. 1980. Habitat prefaencc andforage cunwlllplion by reindeer and caribou ncar Alk'Jsook,Alasb. Arctic :lnd Alpine Research 12(4):511-529.

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUI\IMARIES51

Factors Associated with Predator Distributions

.Gri~zly bear distributions during the caribou calvingpenod III C;lrly June appe:lred to he inl1uenced bv acombination of faclors including seasonal habit;tselection pallcrns, annual variations in sllOwmelt, andannual diStribution patterns of calving caribou

Within-year (1983-1993) spati:lI,listributio·n patternsl)f radiO-COllared grizzly bears tJid not differ among time

III

IIIII

I

Section 6: Predators

DOllald D. }Ollllg, Thomas R. McCabe, Robert Ambrose.Gerald W Garlwr; Greg 1. Weiler; Harry V. Reyno/d.r,Mark S. Udel'irJ., Dan J. Reed, Ilnti Bnui Griffith

Calving caribou (Rallg/fer ranmdlls) of the CentralArctic Iwrd, Alaska, have avoided the infrastructureassociated with the complex of petroleum developmentareas [rolll Prudhoe Bay 10 Kuparuk (Cameron CI a!. 1992,Ncllcmann and Cameron 1998, and Section 4 of thisdocumellt). Calving females of the Porcupine caribou herdlllay similarly avoid any oil field roads and pipelinesdeveloped in areas traditionally used during the calvingand post-calving periods. This may displ.1cc the c:lribou(emaIL's 3ml calves to areas cast and sOlJlh of the 1002Area of the Arctic National Wildlife Refuge.

Increased calf mortality could occur if calving caribouarc di~placed into areas that have a higher density ofpredators, higher rates of predation, or where a higherproportion of the predators regularly usc caribou as a foodsource (Whillcn et,ll. 1992).

Qur study assessed predation risks to caribou calvingin the 1002 Area versus calVing in potential displacementareas. Due to funding constraitlls, our research focuscd ongrizzly bears (UrSlls ilretos), with wolves (CallIIS II/pus)and golden eagles (:1quila ehr)'S{1l!tos) receiving onlycursory aUention. Our research objectives were I) tot:ollljJ;lrC rcla tive ,lbundancc \} f prCtl.ltors within the 1002Area with thai in adjacent periphcml areas, 2) todetermine fat:tors :Jffecting predator abundance on thecalving grounds, alld 3) to quantify the \lSl: of caribou as afood source for predators and the importance of caribou tothe producti vity of predator populations using the coa.,talplain of the Arctic National Wildlife Refugl:.

To al·curately describe the activities ofpretJatorsrdativc to calving caribou, We divided the study area into3 naturally occurring physiographic zones: wastal plain,which included Virtllally all of the 1002 Area « 300 III

elevation); foothills (301-900 1lI ek"ation); and mountains(> 900 III elevation).

Landscape II.~C distriblilions were estimated with fixcd­kernel analyses ll.~ing Least Squares Cross Validation(Silvcnnan 1986, Seaman et al. 1996, 1998, 1999).COIlce!ltra{Ni liSt' art'as were defined as the utilizationcontour that included sites with gr~'ater than averagedensity (Seaman et;ll. 1998). In all cases, sampling waslimited to the llonh .,Iope of the Brooks Range.

Eagle distribution estimates were based on aerialStlrvey locations of 202 nest structures that \I'ere no doserthan I km from adjacent stntctures. Wolf distribtlliollestimates wcre based on al:rial survey locations of 22 densill the Arctic Rclitgc and Ilnrthwcstcm Yukon lcrritorv,Canada. Addition.ll \Volf dens in the foothills and .illlmnlains to the ea.'t of tlle l'stimated wolf t:oncelltr.tlcd

use area probably existed, but were not documented in theanalyzed data set.

Grizzly bear distributions were estimated annually,based on 23-60 annual locations of radio-collared bearsduring the first week of June, 1983-1994. No grizzlybears were radio-collared in Canada. Grizzly bear habitatuse was investigated using Chi-square tests (Neu ela!. 1974). Distance-based tests of independence (Diggkand Cox 1983) as well as analysis of variance procedureswere used to COmpare grizzly bear and calving cariboudistributions.

Predator Distributions

Predators (grizzly be,lrs, wolves, and nesting goldeneagles) in general were more abundallt in 11le foothills andJllountains than on the coastnl plain (Fig. 6.1). Thedistribution of grizzly bear radio-locations relative to thecoastal plain, taothill, and mountain zones was 11011­f:1lldOIll (P < 0.0001, Chi-square).

In all years, the foothills received greater usc by bearsthan expected, whereas the coastal plain received less usethan expected (P < 0.05), except in 1990 when the coastalplain was used in proportion to its availability. Wehypothesize that bears were more :lbundant in thefoothills because the rolling hills proVided greaterdiversity in topogr3phy, vegetation, and phenology thanthe l1atter coastal plain. Other studies have reported lowergrizzly hear densities 011 the arctic coastal plain th:m inthe fOOlhills of the Brooks Range (Miller et a1. 1997.Reynolds 1979).

Radio-collared wolves were more likely to be found inthe foothills (55%) and mountains (36%) Ulan on thecoastal plain (9%). All active wolf dens (n = 11) werelocated in the 1ll01llltnins, with the exception of one denlocated ill the foothills. Since 1982, there havc been lIOreponed cases of wolf deliS on the coastal plain of theArctic Refuge.

All 170 golden cagle nest SlnlCIUrcs, including 22active ncst sites, that were located within 30 km of the1002 Area were found in thl: foothills and mountains(Young et al. 1995). Subadult golden eagles, howl:ver,were ,1bundant on the Arctic Refuge coast31 plain andfoothills where their distribution.~ coincided with those ofcalving c:lfibou.

5~ BlOLOGICAL SCIENCE REPORT USGSII3RD 2()(J2·(){){lJ

Figure 6.1. Distribution of a) golden eagle (Aquila cllfysaeloS) nest structures, b) wolf (Canis lupus) dens, and c)griZZly bears (Ursus arc/os) near the calving grounds of the Porcupine caribou herd. Solid yellow lines encloseconcentrated use areas (CUA. siles with grealer lhan average observation density). solid while lines delineate 99%use distn'bulions (UD), and the dashed red line delineates the approximate 300-m·elevalion boundary between thecoastat plain and foothill/mountain physiographic lones. The ouler pen'meters of all annual griZZly bear fixed kernelestimates of CUA and 99% UD are depicted.

period~, whereas concurrent distributions of cal\'ingcaribou did differ. This suggests that annual grialy beartlislriblllions were inOuenced less by the distribution ofcalving caribou than by other factors (e.g., annualsnowmelt pallerns). Among-year differences (1' < 0.05) ingrialy bear spatial distribution panerJls suggest that;\lInual variations in snowmelt contribute 10 annual beardistribution patterns.

Radio·collarcd grizzly bears were relocated morefrequently on the coastal plain in years when snowmeltoccurred early (38.9%) or normally (23.S'ib), ,IS in 1990

and 1989, respectively. than in years when snowmeltoccurred late (12.1'.',,). as in 1988. Distributions of radio·collared bears and caribou cows with calves tended to bepositively associated in 1988 :l11d 1989 (i.e .. years of lateand normal snowmclt. rcspectively, when calvingoccurrcd primarily in the foothills). and negativelyassociated in 1990 (i.e.. a year of carly snowmelt whencalving occurred primarily O!1 the coastal plain) (Young eta1. 19(4).

Analyses of concurrent grizzly bear and calvingcaribou distributions in 1983-1993 indicated that bears

Il:

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL W1LDLIFE RESEARCH SUl\·lMARIES 53

selected high or medium caribou density zones in 5 of 9(56%) years, but avoided the highest density of caribou in2 (22%) years. Two ye:lrs were not comparable.

During the caribou calving period, radio-collaredwolves were located primarily in the mountains andfoothills where their activity was associated with densitcs. All known wolf dcn sites on thc North Slope of theArctic Refuge have been locatcd in the mountains andfoothills. Thus, the avai labilily of suilable: den sitc.~

appears to be the primary factor innuencing wolfdistributions during the calving period.

Factors affecting the distribution of nesting goldeneaglcs differed from those of subadult birds. Nesting oradult birds sought suitable nesting habitat on cliffs foundprimarily in the foothills alld mounlains in proximity 10

colonies of Arctic ground squirrels (Spermophilus parry's),their primary prey (Young et al. 1995). Subadult birdsappeared to be associated primarily with distributions ofcalving caribou.

Rates of predation

Use of caribou as a food source varied among andwithin predator species. Of 26 gri7.Zly bear observationsurveys that were stlccessfully completed (>1 hr), 8 (31%)included a kill of a caribou calf (Young and l\kCabe1997). Kill rates of caribou calves ranged from 1.0 10 6.3killslbear unit/day; a bear unit being a solitary individual,a family group, or a male with I or more consorts.

Trends in the data suggested that bears were morelikely to encoullIer and kill caribou calVes as calvingdensity decreased. This suggests that predator swampingmay be an effective anti-predator strategy by calvingfemales of the Porcupine caribou herd with respect topredation by grizzly bears.

Radio·collarcd wolves were relocated in the vicinityof caribou 34% of the time and on caribou carcasses 9%of the time. Productivity was similar (P > 0.05) between 3lVolf packs with access (4.3 pupsflilter) and I packwithout access (4.2 pups/litter) to the traditional cariboucalving grounds. Because there are few wolves (20-40)and their distributions are usually separated from tho.'ie ofcalving caribou, wolves kilt relatively few caribou duringthe calving period.

Based Oil prey remains collected at nest sites, 1988­1990, we observed little evidence of use of caribou bynesting golden cagles. Ground squirrels were theirpredominant prey (Young et at. 19(5). Subadult birds,however, arc important pretbtOfs of cal\'illg caribou(Whitten ct at. 1992; sa (lim sCcriVlI 3 (lIth;s rqwrr).

References

Cameron, R. D., D. J. Reed, J. R. Dau. ;lnd W. T. Smith. 1992.Redistribution of calving caribou in response to oil fielddevelopment on the Arctic Slope of Al;lska. Arctic 45:338­342.

Diggle, P. J., ;lnd T. F. Cox. 1983. Some distance-based tests ofindependence for sparsely-sampled multivariate spatialpoint patterns. International Statistical Review 51: 11-23.

Miller, S.D., G. C. White, R. A. Sellers, II. V. Reynolds, J. W.Schoen, K.Titus, V. G. Barnes Jr., R. B. Smith, R. R.Nelson, W. B. B(\lIard, ;"Ind C. C. Schwartz. 1997. Brownand hlilek bear density estimation in Alaska usingradiotelemetry and replicated mark-resighttechniques.Wildlife Monographs 133.

Nelkmalln. c., and R. D. Cameron. 1998. Cumulative impactsof an evolving oil-field complex on the distribution ofcalving caribou. Canadi;ln Journal of Zoology 76: 1425­1430.

Nell, C. W., C. R. Beyers, J. M. Peck, and V. Boy. 1974. Ateehniquc for analysis of Iltilization-av:li!ability data.Journal of Wildlife !'.bnagemcllt 38:541-545.

Reynolds, 1-1. V. 1979. Population biology, 1lI0\·ement,distribution (\nd h;lbit;lt utilization of a grizzly bearpopul:ltion in NPR-A Pages 129-182 in Work Group 3,Studies of selected wildlife ;lnd fish and their use of habit:ltson the l(\nds in and (\djaeent to tlle National PetroleumReserve in Alaska 1977-1978. u.s. Department of theInterior Niltional Pelroleum Reserve in Alask:l 105(c) LandUsc SlIldy, Anchorage, Alaska, USA.

Seaman, D. E., B. Griffilh, and R. A. Powe]1. 1998.KERNELHR: a program for cstimming ;lnimal homer:lnges. Wildlife Society Bull<'tin 26:95·100.

__, J. J. Millspaugh, B. J. Kernohan, G. C. Bnlllllige, K. J.Raedeke, and R. A. Gillen. 1999. Effects of s;lmplc size onkernel home r;lnge estimates. Journal of WildlifeManagemcnt03:739-7.n.

__, and R. A. Powell. 1996. An eV;llu:ltion of the accuracyof kernel density cstiJ1);ltors for home rangc analysis.Ecology 77(7)2075-2085.

Silverman, B. W. 1986. Density estinwtion for statistics anddata analysis. Ch(\pman and Ilall, Lond()J1, Englilnd, UnitedKingdom.

Whitten. K. R., G. W. Garner. F. J. M:Hwr, and R. B. Harris.1992. Productivity and early calf survival in the Porcupinecaribou herd. Journal of Wildlife M:\Ililgel11ent 56:201-212.

Young, D. D., and T. R. McC;liJe. 1997. Grizzl)' bcar predationratc_~ on carihou calvc$ in northeastern Alaska. Joumal ofWildlife Managelllclll 01 (.l): I056- 1060.

___'_~, G. W. Garner. and Il. V. Reynolds, 199·1. U,.., of:1 Jisl:mce-oased lcst of irldcpemknce to [ll..,asurc brownIlc;lr·cariblJu as,odalion in norlheaSl('rIl .-\I:lska. NinthInlernational Confercnce of ll('ar RcsC;1fch !'.lanagemelll<):-135--1·12.

__, C. L McllllYh~. P. J. Benle. 1'. R. l\lcCahe :mu !{. E.Amhrosc. 1995. N..,sting by goltlen eagles on lhe NorthSlop" of the Brooks Rangc in north<':\stcrn Alask:l. Journal01 Field Ornithology 66(3):373-379.

54 BIOLOGICAL SCIENCE REPORT USGSII3RD 2002-0001

Figure 7.1. Number of muskoxen observed in the regions firstoccupied after reintroduction - the 1002 Area. Arctic National WildlifeRefuge, Alaska, USA, during spring censuses, 1982-2001

(95% contour) were used to document changes indistribution and range expansion over time.

Locations of muskoxen sc:en during seasonal muskoxsurveys in the Arctic Refuge, as well as locations ofmixed-sex groups seen during other studies, were used todocuillent the Continued expansion of the populationdistribution fromI994-200J. Reconstructed models of thepopulation estimating maximulll and predicted populationgrowth were used to evaluate how changes in calfproduction, survival, and emigration affected localabundance of muskoxen.

The Arctic Refuge's reestablished population ofmuskoxen grew slowly for a few years and then increasedrapidly for more than a decade (Reynolds 19980).Between 1977 and 1981, the population grew at r = 0.24,a rate approaching nlilximum growth. In 1986,368muskoxen were counted in the 1002 Area but after 1986numbers of muskoxen declined (Fig.7.1). The r:\te of 'increase slowed to 0.14 in 1982-1986 and to <0.00 in1987-1995.

11le rme of population growth (r =0.14) in 1972-1996(Reynolds 19980) was similar to rates recorded for othere;o;panding populfltions of lIluskoxen in Alaska andCanada (Spencer and Lensink 1970, Gunn et al. 1991).An introduced population of llIuskoxen in Greenland wasstill imlpting 25 years after release, but those animalswere in an area of abundant high-quality forage with nopredation and low snowfall (Olesen 1993).

ln 1996-2001, numbers of muskoxen counted in the1002 Area ranged from 16810 212 (P. E. Reynolds, U.S.Fish and Wildlife Service, unpublished data) and indicatethat muskox abundance is still declining slowly (Fig. 7.1).

Factors affecting changes in the number of muskoxenin the 1002 Area of the ArctiC Refuge included changes inrates of sllccessful reproduction and sun'ival as well nschanges in animal distribution.

Calf production was the only source of increase in thisreestablished population; a lack of adjaccnt populations

Section 7: Muskoxen

Patricia E. Reynolds, Kenneth J. lVi/SOli, and David R.Klein

Dynamics and Range Expansion of aReestablished Muskox Population

Muskoxen (O~'ibos moschatlls) disappeared fromAlaska in the late 1800s, but relUrned to the ArcticNational Wildlife Refuge when animals werereestablished into areas of former range in 1969- I970(Klein 1988). Released at Barter Island (Kaktovik) andthe Kavik River, Illuskoxen initially moved into regionsthat encompassed the 1002 Area on the coastal plain ofthe Arctic Refuge. From 1974 to 1986 the muskoxpopulation grew rapidly. By 1987, however, numbersdeclined in the regions that they had first occupied(Reynolds 19980).

Petroleum e;o;ploration and develoJlment could occur illlllusko;o; habitat in the 1002 Area of the Arctic Refuge.Status of the muskox population and factors related totrends in local abundance need to be detennined ifchanges resulting from natural processes arc to beseparated from those that might result if industrialdevelopment is permitted in the Arctic Refuge.

We developed a study with the following objectives to

understand the dynamics of the muskox population in andncar the 1002 Area of the Arctic Refuge: I) detenJlineabundance and rates of population increase, production,and survival; 2) document changes in populationdistribution over time; and 3) evaluate factors associatedwith changes in the number of muskoxen.

Numbers of muskoxen seen during annual censuses in1982-200 I, combined with data from earlier studies, wereIlsed to estimate animal abundance and population tn::ndsof n1U.~ko:"en in the 1002 Area (regions first oecupil.'tl)and adjacent areas to the cast and west (regions occupiedlater) (Reynolds 1998(/). Rates of.mcccssful calfproducTion (defined as calves: 100 femules >2 years oldpresent in late June), survival of calves and ye;lrlings, andlong term reproduction pattem.~ by marked femalellIuskoxen were determined from annual sex and ageeOlllposi tion counts nl.1dc from Ihe ground in 1983-200 I.

fbdio-collared auult muskoxen were relocated (j

tillles!ye;lr from 1982 to 199·110 determine seasonal andanllual variability in pOJlulalioll distribution and todoculllent adult lllonalities. Locations Were determinedfrom the air with a global positioning system (GPS) orwere plotted on 1:63,360-sealc llWpS. The adapti\'e-kt~rnd

technique within the computer program CALHOl\IE (Kieet a1.19(6) was llsed to delineate the size and locations ofregions used by mixed-sex groups of llluskoxen in 1969­1981,1982-1985,1986-1989. and ]l)90-199J. Core a/"ell.\"

OlliSi' (70% adaptive-kernel contour) and !01II1 range

87.11nlU

91 93 95 97 [19 01YEAR

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMt1ARIES 55

of muskoxen precluded immigration. Growth of thepopulation during its most rapid increase was due to highrates of reproduction and survival of muskoxen in newlyoccupied habitats. In the 1002 Area, indices of successfulcalf production reached a maximum between 1977 and1980 (87 calves: 100 females >2 years old) during yearswhen successful reproduction by some 2-year-old femalesand 100% reproduction among females >2 years oldoccurred in some groups (Jingfors and Klein 1982).

Calf production declined between 1983 and 2001 (Fig.7.2). In 1983-1986, as the rate of population increasebegun to slow, calf production in the 1002 Area averaged61 calves: 100 femaks >2 years old compared with 49 in1987-1990,41 in 1991-1994, and 28 in 1995-1999(Reynolds 1999). In June 2000 and 2001, very few calvcs«5 calves per 100 females >2 years old) were seen.Because calves were counted several weeks after birth,we could not determine if changes in the production ofmuskox calves were due to lower fecundity or increasedneonatal mortalities.

Reproductive patteOls of radio-collared females in the1002 Area showed similar trend.~. Mean reproducliI'('intervals (number of years between successfulreproductive event.~ in a 3-year period) increasedsignificantly (r = 0.95, II = 6, P = 0.0009) between 1982and 1999. By 1991-1993, 1Il0st marked femalessuccessfully reproduced at intervals of 210 3 years, ratherthan every year (Reynolds 200 1). Percentages of markedfemales without calves for 3 or more consecutive yearswere 0% in 1982-1987, 15% in 1988-1990, and 25% in1994-1996 (Reynolds 2001). In summer, body weights of8 lactating muskoxen (me:lIl = 223 kg, range = 188~254

kg) were not different (I =2.2, df = 10, P =0.167) from Snon-lactating females (mean = 196 kg, range = 136-254kg) and were similar to weight.~ of female muskoxen inother wild and captive populations (Reynolds andReynolds 1999).

Unlike calf production, which declined in 1983-1995(r= 0.75, P < 0.001), calf and yearling survival did notdecline over time (calf survival: r =0.01, I' =0.71;yearling survival: r =0.01, P = 0.82). Annual variabilityin young animal survival followed the same annual trendsas calf production and was related 10 SIIOW depth and thelength of the snow season (Reynolds 1998a).

Between 1983 and 1999 the percelllages of radio­collared muskoxen dying each year were variable butshowed an increased trend (Reynolds 1999). Sources ofmortality included kills by predators, including humans.

Legal hUllling of muskoxen in the Arctic Refuge beganin 1982. Over lime, the lIumber of permits issued ctlchyear increased frolll 5 males only to 12 males and 3femaIcs. The season was expanded from 2 to 8.5 months.An :l\'erage of 7 muskoxen was killed in 1985-1989compared with an average of 10 in 1995-1999. About 3%of the muskox population in the Arctic Refuge washarvested annually from 1990-1999 (P. E. Reynolds, U.S.Fish and Wildlife Service, unpublished data).

Kills or scavenging of muskoxen by grinly bears(Ursus arcros) in and near the Arctic Refuge increasedsignificantly between 1986 :llld 2001 (/3 =0.504. df= 18,I' < 0.001) (Fig. 7. 3). Known kills of muskoxen bygrizzly hears ranged from 0-2 deaths per year before1993, 1-4 de:lIhs per year in 1993-1996, and 5·10 deathsper year in 1997-2001 (Reynolds et al. 2002).

Forty-seven deaths of adult or sub-adult muskoxenfrom known grizzly bear predatioll occurred between1982-2001. Of these, 28 nll1st..:oxen died during 10incidellts of multiple kills in which bears killed more thanone Illuskox from a group. r-.lost of these kills (79%) lOokplace between May 1998 and June 2001 (Reynolds et al.2002).

Gri7.7.ly bears likely also killed Illuskox calvcs andcau.~cd other mortal itics of young calves that were

DMu;p"'~J~'

.Sir.....,kilos

• P""ibl<! ~ih Of .ca"'fIo;)'r>J

Y"~f

Figure 7.3. Number of mus~oxen killed or scavenged by grizzly bearsfrom April 1982lhrollgh June 2001 in northeastern Alas~a. USA. (fromReynolds et al. 2002)

•

••

••

y~.3.13JJ,.71J~'~

R'" () 7\38

••" ro~

"," ~0e

",0

""0,",

15 ","0,,"z

~ g ~ § ~ ~ ~ ~ g gg g ~ ~ ~ ~ ~ ~ ~YC~f

Figura 7.2. Changes in rales 01 successful production of mus~oxcalves in the regions first occupied aller reintroducliOrl-the 1002Area. Arctic National \Vildlife Refuge. Alas~a, 1983-2001. Successfulcall production was measured by countirlg lhe number of calves per100 females >2 years of age in lale June.

56 BIOLOGICAL SCIENCE REPORT USGSfBRD 2002-0001

descrted during predalion cvcnts. ~lultiplc kills of calvcswere observed in C:lI\ada (Clarkson and Liepins 1993).The increase in kills by grizzly bears suggests thaiprcdlllion may have bccn onc factor that resulted in vcrylow numbers of calvcs in late June of 2000 and 200J.Deep snow and a prolonged winter seaSon in 2000 and2001 also likely contributed to the low numbers of calvesseen in Ihose ycars and lllay havc cxacerbated the numberof predation events (Reynolds et a!. 2002).

Shifts in distribution :lnd emigration also affectednumbers of muskoxen in the 1002 Area of the ArcticRefuge. Following their release in 1969 and 1970, mostmuskoxen became associated with 1 of 3 mixed-sexgroups in 3 regions of the Arctic Refuge. The regions firstoccupied were located between the C3nning and AichilikRivers within the boundaries of the 1002 Area (Fig. 7.4a)(Reynolds 1998a). After 1986, muskoxen in mixed-sexgroups colonized new regions cast and west oftlle 1002Area (Fig. 7.4c,d) (Reynolds 1998a).

In 1995, about 800 llluskoxen werc counted in theenlire range of the population (Table 7.1), which hadexpanded westward to the Itkillik River, Alaska, andeastward to the Dabbage River in northem YukonTerritory, Canada (Reynolds 1998a). In 1998-2001,mixed-sex groups of muskoxen continued to expand theirrange west to the ColviJle River, southwest along theSagavanirktok River, and south and cast of the BabbageRiver in northwestcrtl Canada. During these years, <700muskoxen were counted throughout the total range of thepopulation (Table 7.1) (P. E. RCyllolds, U.S. Fish andWildlife Service, unpublished data, E. A. Lenart, AlaskaDcpartment of Fish and Game, unpublished data, and D.A. Cooley, Yukon Renewahle Resources, unpublisheddata).

Differences betwecn the observed and predictedabundance in the 1002 Area, based on rc-collstmctedpopulation projections. suggest that ch:\llges in muskoxcalf prodllction and animal survival caused most of thedecline ill the rate of population growth (Reynolds 1999).Density dependcnt factors as well as annual variability insnowfall and increasing rates of predation atllikelyinfluenced observed changes in calf production andanimal.wryival.

Emigration of mixed-sex groups of muskoxen alsoreduced the number of muskoxen in the 1002 Area(ReynoldS 1998/1, 1999). In 2000 and 2001, the additionalemigration \)f mixed-sex groups containing markedanimals ;lJld the low rates of successful calf production«5 calves per 100 females> 2 years old) contributed tothe declining trend in Ilumhers of muskoxen in the 1002Area (P. E. Reynolds, U.S. Fish and Wildlife Scrvice,unpublished data).

Although muskoxen arc continuing to cKpalld intotheir fonner range ill northem Alaska and nortllwestcO\Canada, numhers of IllllskllXen inlhe ]()02 Area arc llOt

likely to increase from their present leyel of <250 animalsin the ncar future.

If exploration and extraction of petroleum resourcesarc permitted in the Arctic Refuge coastal plain,associated industrial activities could further reduce thenumber of muskoxen in the IDOl Area either throughinduced disJlersal or decreased productivity and survival./l.Iuskoxen arc year-round residents of the ID02 area,which heightens their vulnerability. In addition, theirsmall numbers make it less likely that the Illuskoxen canrecover from perturbations.

Status :llld distribution of muskoxen in and near theID02 Area should continue to be monitored to documentfuture trends.

Seasonal Strategies of Muskoxen: Distribution,Habitats, and Activity Patterns

Seasonal shifts in distribution, habitat usc, 3nd activityare means by which animals maximize energy intake andavoid conditions that risk their survival. The muskox is :111

energetically conservative species (Klein 1992) and itsseasonal habitat usc and energy budgets influence itsreproduction and survival (While et al. 1989). Limitedforage availability and energy constraints in winter aswell as potential cumulative effects of disturbancecontribute to its susceptibility.

As year~round rcsidents of the coastal plain of IheArctic Refuge, muskoxen arc vulnerable to humanactivitieS in both winter and SUlllmer. Information isneeded about their seasonal patterns of distribution andactivity to cvaluatc and minimize. potential effcctsassociated with oil and gas exploration and developmelllproposed for the Arctic Refuge's 1002 Area.

Our study to detefllline seasonal patterns of lllusko;"enon the coastal plain of the Arctic Refuge had thefollowing Objectives: I) comp:lfe distribution and habitatusc of muskoxen in different seasons, and 2) deteollineseasonal movements and activity patterns of muskoxen.

rn the Arctic Refuge. snow is present from 8-9 monthse:lch year (September - 1\1ay). Fi\'c seasons werc delinedfor muskoxen based on ecological and hiologicalconditions; call'illg (late l\farch to mid-June), summa(late June to mid-Septembcr), {'arfy winrl'r (lateScptcmber 10 lllid-Nov~l1lber), mitt-wiTlfer (late Novemberto mid-January), and 111ft' will/a (late January to mid­I\farch).

To idemify pt1pulation distribution in different.\caSOJlS, 19-25 radio-collared Illuskoxcn were monitoredand 4 to 6 radio-relocation surveys were flown each yearfrom 1932 to 1995. Locatil)lls of groups of muskoxen,both marked and \lnJ1lark~d, were determined using aglobal positioning systcm or wer~ plotted on 1:63,360sL:;l!e maps.

c) 1986-89

!I

ARCrIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SU~H..1ARIES

b) 1982-85Barter Island

'\~Sagavanirktok R.

r~ Canning R. Barter IslandF.r ---..c::r-"'-0-. ~i:/- J:&..... ........""t..........,

11 .,'. !.i!!'~ ,'";.7 -y'~./'~ I • . :\.~Kongakut R.~ .... ~*_'~"" t',·~\~\ • t ~ ~~Il.:f :I~l.* .::::.. ' ;.-:' ,l...·:· 'F--'~

\ ... . ., • ..•• 'I • • "ir. 'l"~c-a-n-a/d~a-1 \,- " • t-; -' .. ' ..........

(" >'---~ ) )d) 1990-93

rJ-;,r~-:~=""c~a;~;n:[ngR. Barter Island

p, "" ~If// , ,C,), 1(1;;;'.... A",- .r'''''' "(. .~: (~.~ '~". ": t:- -:: '.. # Kongakut R.

(

Il ... ' ..' ....<.~.; .' '~." .. -. : :.-:

~\:., ••_ ,~", •• :., • -, .... -" .. ' ...... f'

;, ......,. 1 '. '~'. £..-' ..:J"'~_~-• '.,. " : '. •••••••••• I

",: ·'1 I'; : .I'y~ .,~:""----~t- ".,',': . Canadar ('ARCTIC NAT10N~E:~~':E REFUG ~.

.. '~ CORE AREAS

\~, 0 TOTAL RANGE

+ LOCATION OF MIXED·SEX GROUP

Figure 7.4. Range expansion of muskoxen in mixed·sex groups in and near the Arctic National WildlifeRefuge. Alaska, 1969·1993. Total ranges were defmed by 95% adaptive kernel contours. Core areas weredefined by 70% adaptive kernel contours. (from Reynolds 1998a)

57

58 BIOLOGICAL SCIENCE REPORT USGSIBRD 2002-0001

Table 7.1. Number of muskoxen seen in different regions in northeastern Alaska, USA, and northweslern Canada in 1982·2000 during pre­calving surveys. GMU 268 and 26C are Slale of Alaska game management units. The muskox population originated from animals releasedadjaC€nt 10 the Arctic National Wildlife Refuge, Alaska. in 1969 and 1970. The muskoxen began to disperse inlo new regions east and west ofthe Arclic Refuge by 1986 (Reynolds 1998a).

)

West altho Arctic 1002 Area in the ArcticRefuge: Itkillik River to Refuge: All Arctic Refuge:Canning Rlyer (GMU canning RIver to Canning River to Canada

Year 26B)0 Alchllik Aiverb (GMU 26C)

1982 219 219

1986 9 386 399

1990 122 273 332

1995 330 228 321

1998 207 213 331

2000 277 189 246

East of the ArcticRefugo in northern Arctic AefugeYukon Terrilory, + west + enslCanadac

2'9

23 431

41 495

146 797

116 654

149 672

\a data source: E. A Lenart, Alaska Departmenl of Fish and Gamo, Fairbanks, Alaska, USAb regions first occuploo; numbers included in Arctic Reruga numbersc dlltll source: D. Cooley, Department of Renewable Rosources, Yukon Territories, Cannda

Figure 7.5. Seasonal changes in rales of mo;'emenl and activitycounts of satellite·collared female muskoxen in and near lhe ArcticNational Wildlife Refuge, Alaska, i986·1992. (from Reynolds 1998b)

4

'.5~, 0~

2.5 •~2 •" ~

B~

05

0

• ... 'AC~\Ilt)'mdex

1\...., .. -.-I.'ovement rate. .. ., ,

, ,, . ._0· --.:: ..... -.•..•..'J ','0" •. '.

'-.......~----.../

"~

20c'E 15•~~ 10,0u ,

o ~~.~~~-,-~-~!.lay Jul Scp Nov Jan Mar

Month

core areas was highly variable in sumlller, out meansdiffered by almost an ordcr of magnitude betweensummer and other seasons. 111e minimum sii'.c of coreareas used in summer wa.~ >4 times larger than minimumcore areas occupied in winter or calving.

rvtuskoxen wac conservative in their daily movementsthroughout the ycar. Most (95%, Il = 2314) lllovelllelllSmade by satellitt-collared llluskoxen wtrc <5 km/day(Reynolds 1998b). Of these, 46<;·" were <1 kill/day.Moderate 1Il0V~ll1entsof 5-10 kill/day took pl3ce primarilybetween June and September (77 of lOS). Only 1 «1%)moderate 1ll0\'ellle11l of 5-10 km/day was recordedbetween January and April. Movement rates >!O kill/day,resulting in relatively long moves, w~re rare (18 of 2314);16 (89%) movements >10 km/day occurred in July.

Mean daily movemenls in summer (2.6 km/day) weregreater (P < 0.05) lhan in other seaSOllS (1.1-1.4 km/day){Reynolds 1998b). l\lcan rales of movcmenl weresignific:Hllly higher (P < 0.05) in July than in olhermonths (Fig. 7.5). Activity counts/minute from satellite-

Seasonal distribution and movemenl rates weredetermined from 15 female muskoxen filled with satcllitccollars (ultra-high frequency platform transmitterteffilinals that were relocated by satellite) (Reynolds1989). Three to.'5 animals carrying satellitc collars weremonitored yearly from October 1986 through MarchJ992. These collars transmitted infonnation about animallocation and activity every second or thinl day for 6 hr/day (Reynolds 1998b).

Seasonal distribution of the population and seasonalhome ranges of satellite-collared mu.skoxen \\"ere'delineated with an adaptive-kerncltechnique and theprogram CALHOl\!E (Kie et :II. 1996). Seasonaldifferences in population distriblllion were cOlllpared asthe overlap of core areas (70% cOlllour), distancesbet\\,'een core-area centers, and core-area sizes.

Mean movement rates (km/day) for ench season andeaeh month were calculated from distanccs moved bysatellite-collared lllllskoxen. Distances were calculiltedbetween consecutive locations at 40-55 hr intervals. fo.1canactivity indices for each season and each month werederived. AClivity counts from 5 satellite-collaredIlluskoxen with >10 days of activity COUlltS pcr monlhwere used to estimate mean aClivity (Reynolds 1998h).

Lmd-cover and terrain lypes, extracted from a land­cover lllap derived from Landsat-Thematic l\lapper dala(Jorgenson et al. 1994), were Ilsed t\) determine seasonaldifferences in habitat usc at a landscape scale. Sekctionralios of 6 laml·cover classes and 5 telTaln typcs wcrebased on prop()l1iol1s present in core :\reas (habitats Ibcd)divided by propo11ions in the entire study area (habitatsavailable) (Reynolds 199811).

The average size \)f core areas IlsL'd by lIluskoxencarrying s:llellilc l:ollars was significmllly larger (Jl <0.05) in SUlllllll:r (223 km') than in calving season or lhe 3wintl:r seasons (27-70 kill') (Rcynolds ] ~)\)ob). Thc size of

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUMMARIES 59

collared muskoxen were also greater in summer (I' <n.OOI) than in otln:r seasons. Activity COUnts differedamong months (I' = D.OOI) and were highest in July andlowest in April during the onset of the calving se;lson.

Seasonal home ranges occupied by females withsatellite-collars overlapped less (I' = 0.01) betweencalving and summer than betweell early winter and mid­winter and betweell mid-winter and late wintCf (Reynolds1998b). This rdlccted Ihe sedentary nature and smallhome range size of the muskoxen in winter. Distancesbetween seasonal home ranges were also small. Thedistribution of the population of muskoxen occupying thecoastal plain of the Arctic Refuge showed lillie changebetween seasons.

AI a landscape scale, muskoxen llsed rip;trian coveralong river corridors, floodplains. and foothills in allseasons. Moist sedge was selected in late winter andcalving; tussock tundra was avoided in late winter. Wetsedge was used in proportion to availability in summer

1002 Area

Arctic National Wildllife Refuge

and early winter bUi avoided in other se;lsons. Uplandshrub was selected only during the calving season andavoided in other seasons. Bare cover (including bareground, water, and ice) was .~eleeted in all seasolls exceptspring. Mountain terrain was avoided in all se;lsons(Reynolds 19981),

Ground-based studies (Wilson 19(2) provided moreinformation at regional and local scales (st'!' 11(':([

subsectioll 011 wimer haln'fm /1.I't'). Locations of mixed-sexgroups of [lluskoxen during summcr and winter surveysdemonstrated the importance of river corridors andadjacent uplands to this population (Fig. 7.6),

The small seasonal shifts in distribUlion and lowmovemcnt ratcs observed in this study con finned thatmuskoxen arc energelically conscrvative throughout theyear (Jingfors 1980, Thing et al. 1987) and that they havca high fidel it)' 10 geographic regions. Seasonal changes inmovcments. activity, and habitat usc were related to

Beaufort SeaBarter Islclnd

r.~nniinn River

60

Kongakut River

120

(/ / Jl

Kilometers

1\) !,

,/

Figure 7.6. Locations of mixed-sex groups of muskoxen seen during winter and summer sUf\leys in the Arctic National Wildlife Refuge, Alaska,USA,1982-1999.

60 BIOLOGICAL SCIENCE REPORT USGSfllRD 2002-0001

I

availability of forage and the energy budgets ofmuskoxen.

In winter, snow limits forage availability and habitatselection (1ingfors 1980). In late winter, muskoxenselected feeding sitc.~ with soft shallow snow(Biddlccomb 1992, Wilson 1992). These siles werefrequently narrow windblown bluffs adjacent to riverswhere snow accumulation was low (Nclkmann ;lUt!

Reynolds 1997), By mid- to late wintcr, riparian willowsand wei-sedge cornllllillitics may be unavaihlblc tomuskoxen as snow depths increase (Wilson 1992, Evansel 31. 1989).

Winter forage of muskoxen is of low {juality (Staal andand Olesen 1992). Graminoids were a dominantcomponent of the latc winter diet of muskoxen inllortheastemAla.sb (O'Brien 1988, Biddlecomb 1992,Wilson 1992). l'vluskoxen, however, can digest low qualitygrarninoids efficicntly and may havc a fasting metabolicrate lower than other ruminants (Adamczewski et al.1994, Lawler 200 I). In winter, muskoxen conserve energyby reducing movements and activity, decreasing the sizeof usc areas, and cOllcelllrating in 11 few habitats whercforalle is not covered with deep snow.

Unlike caribou that calve in early June when nutritiousvegetation is emerging, most muskoxen give birth severalweeks before high quality green forage is available. InAlaska, Dall's slwcp (Ovis dal/i) (Rachlow and Bowyer1994) and moosc (Alces alet's) (Bowycr ct al. 1998) havcsimilar calving strategies to muskoxen. To reproduccsucccssfully, female llluskoxen must bc in good bodycondition at calving timc to fud thc high cos! of lactation.TIleir cnergy-conscrving stratcgy of restrictingmovcmcnts and aClivity and sclecting habitats w.ith lowsnowcovcr allows female muskoxen to Ill:limain bodycondition throughout thc winter and spring (TIling et al.1987).

l\.·lost llIllskoxcn in the Arctic Refuge givc birllJ inApril and l\fay and Iactatc undcr conditions of poorquality forage and harsh wcather. TIdr increased use offoothill terrain and upland shrub during the calvingSC<'lson reflected shifts into areas where snowcover wasshallow or blown free and thcir energetic costs offomging were lower. 11uskoxCll with young calves alsoIllay avoid flooded riparian areas during calving and post­calving periods. Iviovelllent rates of muskoxen carryingsatellite collars rcached a ycarly low in April at the onsetof the calving season.

During thc sllow-free SllTlHner when food quality andIJuantity arc high, muskoxen increase their movel\lcnt andactivity, occupy larger areas, and usc di\'erse habitats asthey forage 011 a variety of high-quality vegetation (Robus1981, O'Brien 19S5). They tr;lck the changing plantphenology in loc:ll arcas to obtain high quality forage andrapidly regain body weight lost during winter. pregnallcy,and carly lactalioll. Musko,xcil that fail 10 regain body

weight arc less likely to breed or successfully reproduce(White et al. 1997) and arc less likely to survivc a severewinter.

In our study we .found that movemcnl mtes :\lldactivity incrcased in June as plants began to leaf out andwere highest in July as live plant biomass peaked (Chapin1983). Movcmcnt rates and activity of muskoxen carT}'ingsatellite-collars began to declinc in August as plantsenescence and rut occurred (Reynolds 1998b).

Seasonal strategies that emphasize energy inlake insUlllmcr and encrgy conservation in winter, combinedwith physical adaptations for cold weather and the abilityto process low quality forage, perrnitlTluskoxen to surviveyear-round in locations seasonally avoided by most otheranimals. Muskoxen arc prcscnt during all scasons in thepotcntial oil exploration and development area of the.·\rctic Rcfuge.

This study did not quantify Ihe effects of pctroleumdevclopment on muskoxen. But human activities thatincrease cncrgetic costs to muskoxcn in winter ordecrease foraging opportunities in SUllllller have thegreatest probability to affc.ct thc Illuskox population.Riparian habitats frequently used by lIluskoxen arc alsolikely to be llsed as sites for gravel and water extractionand wimer road construction if exploration anddevclopment of petroleum resources occur on the coastalplain of the Arctic Refugc.

Exploratory and construction activities in northernAlaska oftentJke placc in winter. Muskoxcn arcparlicularly vulnerable to disturbance in winter because oflimited habitat, thc length of the arctic winter, and theirnecd to conservc cnergy throughout the winter includingthe calving season, The average size of muskox groups islarger in winter than in mid-summer (Reynolds 1993).Large groups of animals often arc more easily disturbcdthan small groups be.cause largc gr0l1ps contain moreindividuals respollsivc 10 pcrturbations.

Effects of human ,Ictivities on muskoxen are likelyrelated to the scale of the activity and the availability ofalternate habitats that can be Ilsed if animals arcdisplaced. lv'Iuskoxen that expanded wcstward from thcArctic Rcfuge usc the widc Sagavanirktok Rivcr valley insummer despite the prescnce of the Daltoll Highway andthe trans-Alaska oil pipdine. Habitats availablc tomuskoxcn in the Arctic Refuge, however, arcgeographically constricted: The coastal plain is narrowerbecausc the mountains of the Brooks Range arc closer tothe l3eilUfort Sea (Fig. 1.1).

If undisturbed, muskoxen generally slay in relativelysmall areas lhroughout the wintcr. Avoidance by industryof these areas used by muskoxen could reduce theprobability of disturbance and displacemcnt of muskoxen.Minimizing human activities in areas occupied byllluskoxcn from mid-winter through the calving seasoncould reduce the likdihood of disturbance during the

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUl\fl\'IARIES 61

'990

60

£ 60

•,.:0

!, 60

t "!00 w

0

FcCdiJ)(l ZOnc

60

".' 60g• "8,

30<a0

"u

I w,00

0

FcC'd;ng ZDnc

Mj~ccnt ZOnc

HlB'

in feeding zones. Non-vegctative cover was greater inadjacent and nonadjacent wnes (Wilson 1992).

Dict selection based on fecal analysis of winter pellets(corrcctcd for digeslibility) indicated a high use of sedges(39.1%) and Illosses (24.6%) (Wilson 1992). Sedges and,grasses were selected (use> availability); and horsetails,lichens, willows, and othcr sl1mbs were avoided (usc <availability). Although selection for grasses was high,grasses did not make up a large proportion of the diet orthe available habitat.

Analysis of mmen samples indicatcd that sedges(31 %), grasscs (19%), mosses (15'."0), and forbs (13%)comprised most of Ihe diet. The proportion of willowswas 8% in millen samples. In other studies, riparianwillows were used by muskoxen in latc winter (O'Bricn1998, Robus 1991). During our study, howcver, snowlimited thc usc of Illost riparian shmb communities.Willows Were browsed in areas where they protmdcdthrough the SIIOW (Wilson 1992).

Snow dcpth was shallower and softer in feeding zonesthan in nonadjacent zones, and shallower in feeding zonesth<lll in adjaccnt zones (Pig. 7.7). Snow depth and

Figure 7.7. Snow depth (em) in muskox feeding zones (areascontaining feeding sites), adjacent zones (unused areas adjacenlloand surrounding feeding zones) and nonadjacenll.ones (unused areas100 meters beyond adjacent zones) in late winter 1989 (Il =20) andi990 (n = 24) on the coastal plain of the Arctic Nalional WildlifeRefuge, Alaska, USA. (from Wilson 1992)

Winter Habitat Use by Muskoxen: Spatial Scalesof Resource Selection

During the snow season, which lasts up to 9 months inthe Arctic Refuge, muskoxen remain in slllall areas,restricted by the availability of forage and by strategiesnccded to conservc energy (Reynolds 1998b). Humandisturbance or destmctioll of their habitat could displaceIlluskoxen from these limited wintering areas.

To detcrmine what kinds of sites arc used byllluskoxen in late-winter and why these sites arc selected,we sct the following research objectives: I) detennineselection of vegetation types based on use andavailability, and 2) compare .~now depth and hardness,vegetation biomass, and environmental variables atfeeding and non-feeding sites. Our study sites werelocated in the eastem half of the Arctic Rduge coastalplain and 1002 Are;l, betwee.n the hgo River and theKongakut River.

Fieldwork was conducted in March, April, and July1989-1990 at 44 late-winler fora,ging sites used bymuskoxen. These sites included/ceding zones (withfeeding microsilcs or craters), llOll-llsul adjacCIlIl.O/lC.\·

(contiguous to the feeding zonc:), and IJOfHISetl

/101Uldjacl'/ll zones (100 III beyond the adjacent zone). Ateach foraging site, a randomly-oriented transect was laidacross the sitc, passing through the center of the feedingwne and defining unused zones.

Foraging sites were located from observations :wdtracks of muskox groups in March and April. Fecal pellctswerc collected for diet selection analysis. Winter foragingsite.~ were relocated the following July when vegctatiolland environmelltal characteristic,~ along transects weremcasurcd (Wilson 1992). Snow conditions, environmcntalcharacteristics, for:1ge cover, and non-vegetatcd coverwcre included in an analysis of variables rel:ued to habitatselection.

In late winter, llluskoxen fed most commonly in moistsedge tundra (37':;') and 1IJS.~m;k sedge tundra (37'7'(,) andIIs(~d these types in proportion to availability. Dryasterrace {9~~.), riparian grass forb gravel bars (7%), wctsedge (5%), partially vegetated tundra (2%), and shmbtundra (2'7c) were selccted Je.~s frequently than theiravailability. Muskoxen were not obser\'5~d feeding inriparian shmh. Dryas ridge, barren ground, or water(Wilson 1992).

Total vegetation cover was gre,lter in feeding zonesthan in unused :ldjacent and non-adjaccnt zoncs. Cover ofevergreen shmbs, sedges, :IflU dead vegetation was grea1er

period when energy conservation is crjticalto survival.Locating permanent facilities away from river corridors,Oood plains, and adjacent uplands could also help toreduce potential effects of industrial development onIlluskoxen.

62 BIOLOGICAL SCIENCE REPORT USGSffiRD 2002·0001

hardness were less in microsites than in unused portionsof feeding sites. Snow depth was the single variable mostinfluential in discriminating between used and non-usedareas. l\luskoxen also appeared to avoid walking throughareas of soft deeJl snow. Most feeding zones were ncarsome type of topographic relief that had been subjected 10wind scarring. Snow depth was shallower in feedingzones of tussock sedge tundra and moist sedge tundra,suggesting that within vegetation types, muskoxen chosefceding zones based on snow depth alone. No differenceswere detected between feeding wnes and adjacent andnonadjacent zones in wetness, slope, micro-relief, oraspect (WilsOll 1992).

Snow depth in fceding zones was deeper in 1989 thanin 1990 (34.5 em versus 23.0 cm, respectively), and IOtalvegetation cover was greater in 1990 than in 1989(Wilson 1992). In 1990, forb cover was greater in feedingzones compared with a greater cover of sedges and non­vcgct:lIive material in 1989 feeding ZOnes. l'....luskoxen didnot select for areas of high IOta I vegetativc cover in 1989,indic<lting that detection of vegetation characteristics mayrequire shallow snow cover. In 1989, mean crater depths(29.7 cm) and mean feeding zone depths (34.5 cm)appro;lched or e:xceeded the miLximum snow depths infeeding areas (>30 cm) observed in other muskox studies(Rapola 1984, Smith 1984).

Partially veget:ded tundra and Dryas terraces had theshallowesl snow; the deepest snow occurred in shrubtundra and moist sedge tundra. Gravel bars with riparianforbs and grass had the greatest total cover of vegetation;moisl sedge tundra had the least. l'....tuskoxen sekctedfeeding zones with shallower snow and greater vegetationCOver compared with what wa.~ available (Wilson 1992).

Winter habitat for muskoxen is limited in quantitybecause animals must select foraging areas with shallowsoft snow and a high cover of vegetation. Areas with liukvegetatioJl or deep hard-packed snow were not used. Inthis study, feeding zones wen~ primarily along narrowbands of windblown vegetated bluffs adjacellt to creeks,rivers, and the coastline, rene-cting the importance oftermin features to habitat sekction (Nellemann andReynolds 1997).

Snow depth was one of the most important variablesdistinguishing used and ullused area in this study ofllIuskox habit:ll (Wilson 1992). Snow tlepth influencl's theavail:lbility of forage and can limit accessibility to someforage types (Evans et al. 1(89). Snow depth affectscllt~rgy budgets. Digging through snow 10 find forage iscncrgl'tiea Ill' enstl)' fllr ungu latc.~ (F,11lCY 1986). The timelost while digging craters also reduccs the daily rate offorage intake Crleischll1an 1988).

In high snow years, when sOllie habitats are notav'lilablc .llld llJuskoxen spend morc l'llergy moving .mdforaging, muskoxen may be energetically constrained,resulting iJllow~r survival and Jess successful

reproduction. As winter progresses :lnd snowaccumulates, or if deep snow falls early in the winter,muskoxen lllay be forced to select foraging areas withdeep snow or low plant biomass.

If muskoxen arc limited in their accumulation of bodyreserves during summer, effects of a severe winter oroveruse of winter range will have greater impacts onreproductive success and survival. If, in addition, animalsare disturbed by human activities and cannot optimallyuse available habit<lts, the effects of a severe winter likelywill be magnified.

Activities associatl'd with the extraction of pctroleumresources on the coastal plain of the Arctic Refuge havethe potential to displace muskoxen into areas of deepersnow wherl' forage avail<lbility is low alld energetic coststo procurc food Me high. Displacement from, or 'permanent lo~s of, limited winter habitat could affectreproductive success and survival of llluskoxen 011 thecoastal plain of the Arctic Refuge.

To minimize potcntial effects of petroleum explorationand development Oil muskoxen in the Arctic Refuge, areasoccupied by muskoxen in winter should be avoided; andareas of potential winter habitat should not be selcr.:tcd assites for pennancnt facilities.

Summary

I\'fuskoxen arc year-round residents of the 1002 Area011 the coastal plain of Arctic National Wildlife Refuge.Number,.; of muskoxen in the Refuge have declined overtime with <300 currently living Oll the coastal plainincluding <250 in the 1002 Area. Calf production has alsodeclined over time.

Severe winters (deep snow and pmlong SIlOW seasons)and increasing rates of predation arc illlport:lIlt factors inthe dynamics of this population. l'vluskoxen haveexpanded their range cast and west of the Arctic Refugecoastal plain and emigration has contributcd to decliningnumbers.

~'fost calves ;If(~ born in April and 1\-lay, several wceksbefore green forage is .1\'aibb1c. To survive the longmonths of winter and to maintain body reserves neededfor successful reproduction, muskoxen conserve energy inwinter by reducing activity and movements. In winter,muskoxen li..'ed on dried sedges .Illd other low qualityforage in areas of low snow. Windblown ridges adjacent10 rivers arc frequelltly used in winter. During the Sh0l1weeks of sumrner, when green forage is available,muskoxc-ll increase their nlOvell1ellts aud activity and feed1)l1 a variety of high quality forage to regain botly \wightbefore the nest winter. River c-olTidors and nearbyuplands are often used by muskoxen in SUllllller.

Muskoxen in the Arctic Refuge arc vulnerable tIldisturbance frolll activities associated witll petroleuillexploration and extraction because of their year-round

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SW...H..,.fAIUES 63

residency, their small population numbers and their needto conserve energy for the 9 months of wimer if they arcto successfully reprodua.

Disturbances that displace lllusko.xen from preferredwinter habitats into areas of deeper snow or that increasetlleir acti vity and 1l10venwnts could significantly increasetheir energetic costs in winter. Female muskoxen that arcrequired to expend grealer energy to survive the winterwilt have fewer reserves for pregnancy and lactation andlIlay not reproduce successfully. Muskoxen frequently usehabitats along or adjacent to rivers. These locations llIaybe sites for gravel and water extraction and winter roadconstruction if petroleum development is permincd in theArctic Refuge.

Avoidance by industry of lucas used by muskoxcn andthe location of permanent facilities away from rivercorridors, flood pl,lins, and adjaccnt uplands could reducethe probability of disturb:lIlce and displacement ofllluskoxen. Status :lIld distribution of muskoxen in andnear the 1002 Area should be monitored 10 doculllcntfuture trends.

References

Ad,llllclewski, J. Z., W. M. Kcrr, E. F. Lmllllcrding, and P. F.Flood. 199·1. Digcstion of low pwtcin grass hay bymuskoxen and callie. JouJllal of Wildlife l\lanagclllcnt58:679-685.

Biddlccomb, M. E. 1992. Comparati\'c p.1l(cms of wimerhabitat usc b)' lllllsko.,cn and caribou in norlhen! Alask:!.lllesis, University of Alaska, Fairbanks, Alaska. USA.

Bowyer, R. T., V. Van Ilallenbcrghe, and J. G. Kic. 1998. Timingand synchrony of parlUrition in AlaskallIllO()Se: long-termvcnus proxim:11 c(fecls of climate. Journal of Manlillalogy79:1332-1344.

Chapin. r. S. 1983. Direct and indircct cffects of t~mpern1Ure on:lrctic pl'lIlls. Polar Jliology 2:47-52.

Clarkson, P. L"'md I. S. Lkpim. 1993. Gri7z1y hear. Ur.wJ()rcIOJ, predation on muskox, ()~'ibl)s 1/l0.fC!latll.f, ~:\lves ncarthe Horton Rher, Northwest Territories. Canadian ridd­NalUralist 107: I00- I 02.

Evans, B. 1'.1., D. A Walker, C. S. Benson, E. A. Nordstrand.:\I1d G. W. Petcrsen. 1989. Spali;1l irllerrclatiom;hips betwccntcrrain, snow distribution :Hld vegetation pallcn!s at an arcticfoothills sile in Alask:!. Hol:lrctic Ecology 12: 270-278.

F.1ncy, S. G. 1986. Daily energy budgcts of caribou: asiJ11ulati()ll appro'lch. Disscrtalion. Uni\'ersity of AI:lska.Fairbanks, Alasb, USA.

Fki5chman. 5, J. 1988. A llwdcl of the cncrgy required hycaribou to dig a fe<.'ding crater, Pag<.' 186 ill. R. D. Callinon.and J. L Davis. editors. Pro(,'c<.'dings \)f the Third NorthAllleric:11l Carihou Worbhop. !\Iaska Dcpartlllcill of Fishand Game Wildlife Technicaillulletin S.

Gllllll, A., C. Shank. and B. MtLcan. 1991. The history. Matus:lIld managemerll of muskoxcn on Bank.> Island. Arctic44:1SS-195.

Jingfors, K. T 19110. Habitat relati'lrlships :lilt! activity paHernsof a r"introduced I11lJsko)( P'>]lul:llion. 'J1lCsis, Universit)' ofAlasJ.::l, F;lirhanks, AI:lska. USA.

__' :lnd D. R. Klein. 1982. Productivity in recentlyeMablished muskox populations in Al:lska. Journal ofWildlife Managclll<.'.nt 46: I092- 1096.

Jorgenson, 1. C., P. 1:. Joria, T. R. McCabe, B. R. Reitz, M. K.Raynolds, M. Emers, M. A. \Vrlms. 1994. Users guide forIhe land·covcr map of the coaslal plain of the ArcticNational Wildlife Refuge. U.S. Fish and Wildlife Scrvice,Fairbanks, Alaska, USA.

Kie, J. G .. J. A. Balwin, and C. J. Evans. 1996. CAU-lOME: aprogram for eSlimating animal home nmges. WildlifeSociety Bulletin 24:342-344.

Klein, D. R. 1988. The eslablishmcnt of muskox populations bytranslOC:ltion. Pagcs 298-317 ill L. Nielsen and R. D.Brown, editors. Translocation of wild a.nimals. WisconsinHumane Society, Inc., ~Iil\\'aukee. Wisconsin, USA.

__. 1992. Comparativc ecological and behavioraladaptations of Ol'ibo$ l/10schalU$ and Rangijer tarandlls.Rangifer 12(2):47-55.

L,l\vler, J. P. 2001. Ueat incremcnt and methane production byilluskoxcn fed browse. Dissertation, University of Alaska,Fairbanks, Al:lska, USA.

Nellemann, c.. and P. E. Reynolds. 1997. Predicting laIc winterdistribution of muskoxen Ilsing an indcx of tcrrainruggedness. Arctic and Alpinc Resean:h 29:33·1-338.

0'Bricn, C. Ii. 1988. Characteri7,ation of muskoJl habit.11 innortheastern Alaska. 1l1csis, University of Alaska,rairbanks, Alaska, USA.

Olesen, C. R. 1993. Rapid population inere:loe in:m introducedllluskox population, West Greenland. Rangifer J3(1);27-3J.

Raehlow, J. L., ,llld R. T. Bowycr. 1994. VariabililY in materna]behavior of DalJ's sheep; environlllentaltracking or adaptive"trateg)'? Joumal of Mamm:llogy 75:328·337.

Rapota. V. V. 1984. Fceding ecology oflhe Taimyr lIlusko;l;en.Pages 75-80 in D. It Klein, R. G. White, and S. Keller,cditors. Proceedings of Ihe First Intcmatioll:ll Musko;l;Symposiulll. Biological Papers, Uni\'Cfsity of Alaska S~cialReport 4.

Reynolds. P. E. 1989. An experimental satellile collar fornluskoxen. Canadian Joum:ll of Zoology 67: I 22-112-1.

_._.1993. 111e dyn:lInic$ of mtlsko;l; social groups innortheastern Alaska. Rangifcr 13:83-89.

__. 1998a. Dynamics and range expansion of:l rcestablishedTlluskm popu!atioll. journal of Wildlife l\lanagement62:734-744.

__. 1998b. Seasonal distribution, activity and h:lbitatuse ofnluskoxen in northcastcrn Alaska. Chaptcr 2 ill Ecology of arcestahlj,hed populatioll of musko.'Cll in northeastcmAl:ISb. Dissertation. University of Alaska. Fairbanks,:\laska. USA.

1999. Factors assm:ialeu with stabililalion of muskoxl1ulllbers ill northeastern Alaska. Paper presented at theTClllh Arctic lIl]gtllat,~ Conference. 9-12 August 1999.Tromso, N()[\,\'~lY. Suhmilled 10 Rangi fcr lin final rc\'i~i(ln).

.__. 200 I. Rcprodllcti\'c pallerns of f,'male llllbko.,cn innortheastern Alask:i. Alees 37(2): 1-8.

_____, :111\1 H. V. Re)'nolds. 1999. Hody weight andrC'pfllducdvc status of felllaic muskoxen in northeaSlcrllAlilSka. P:lper prcsented at thc Telllh Arctic UngulateConference, 9-12 August 1999, Tromso, Norw~l)'. Suhmil1cdto Rallgifcr (in fjnal rcvision).

64 BIOLOGICAL SCIENCE REPORT USGSfBRD 2002-0001

__, __, and R. Shideler. 2002. Predation and multiplekills of muskoxen by grizzly bears. UrsuS 13:in press.

Robus, M. A. 1981. Muskox habitat and use palternS innortheastern Alaska. Thesis, University of Alaska,I~airbanks, Alaska, USA.

Smith, T. IT. 1984, Population status and management ofmuskoxen on Nunivak Island, Alaska. Pages 52-56 ill D. R.Klein, R. G. White, and S, Keller, editors. Proceedings ofthe First Intcrnational Muskox Symposium. BiologicalPapers, University of Alaska Special Repon.1,

Spencer, D. L., and C. J. Lensink. 1970.111C muskox ofNunivak Island. Journal of Wildlife Manage!llelll 34: 1-15.

Slaaland, B., :md C. R. Olesen. 1992, Muskox and caribouadaplation to grazing Oil the AngujaarlOrfiup Nunaa range inWest Greenland. Rangifcr 12:105-115.

'111ing, fl., D. R. Klein, K. Jingfors, and S. Holt. 1987. Ecologyof muskoxen in Jameson Land, northeast Grl'cn land.Holarctit.: Ecolog)l10:95-103.

White, R. 0., D. E Holleman, and B. Tiplady. 1989. Se;lsollalbody Weight, body condition, and lactationaltrcnds inmuskoxen. Canadian Journal of Zoology 67: 1125-1133.

__, J. E. Rowell, and W. E. Hauer. 1997.TI1C role ofnutrition, badycollditian nod bctalioo on calving success inmuskoxen. Journal of Zoology London 243: 13-20.

Wihon, K. J. 1992. Spalial scales of Illusko;r; resource selectionin latc winter. TIlesis, Univer:;ily of Alask;l, Fairbanks,Al.1ska, USA,

ARCflC REFUGE COASTAL PLAIN TERRESTRIAL WILDLiFE RESEARCH SU~HvfARIES 65

Section 8: Polar Bears

SrCW'!l C. Amsrrup

Movements and Population Dynamics of PolarBears

Polar bears (Ursus !I1aritimus) arc hUflled througholUmost of their range. In addition to hunting, polar bears ofthe Beaufort Sea region arc exposed to mineral andpetroleum extraction and rdated hUlllan activities such asshipping, road-building, and seismic testing (Stirling1990).

Little was known at the start of this project about howpolar bears move about in their environment; andalthough it was understood that many bears travel acrosspolitical borders, the boundaries of populations had notbeen delineated (Alilstrup 1986, Amstrup et al. 1986,Amstrup and DeMaster 1988, Gamer et al. 1994, Amstnlp1995. Amstrup et al. 1995, Amstrup 2000).

As human populations increase and demands for polarbears and other :lrctic resources escalate, managers mustknow the sizes and distributions of the polar bearpopulations. Resource managers also need reliableestimates of breeding rates, reproductive intervals, liller.~izes, and survival of young :llld adults.

Our objectives for this research were I) to de!emlinethe seasonal and annual movements of polar b~ilrs in theBeaufort Sea, 2) to define the hound aries of thepopulation(s) ll.~ing this region, 3) tQ detennine the sizeand status of the Beaufort Sea polar bear populatloll, and4) to establish reproduction and survival ratcs (Amstmp2000).

One-hundred-fifty_tllrcc satellile radio collars (P1Ts),fillcd to 106 adult female polar bears in the Beaufort Sea,were relocated 37,277til1les between 19R5 and 1993(Arnstrup 1995, Amstrup 2000, Amstmp ct at. 2000).Polar bears were observed to move more than 4 km/hr forextcnded periods, but mean hourly rates of movementvaried from 0.30-0.96 kill/hr. Females Wilh cubs hadlower hourly rates of movement than females withyearlings and those (single females) without YOllng.

~'fovellient rates varied si~nificall!ly among months:thcy ~enerally were lowcst in sprin~ and late summer andhighest in early winter (Amslmp 1995, Amstrup et al.2000). Geographic displacemcnls frolllthe beginning tothe cnd of each month were slllaller for females with cuhsof thc year than for single females, and larger illNovcmber than in April.

Tn I\by, JUlie, July, and August, radio-colbred bears.~hifted locations to the north. Collared bears moved hackto the south in October. Mean tOlal distances moved c:lchIllonth ranged from 186-492 kill. Total movcments inDecember wcre larger than those measured in April, May,JUly, AuguSl, and September, and lOlalmollthly

movemcnts of females with cubs were lower than singlefemales.

Total annual mOVemcnts ranged from 1,454-6,203 km.Bcars lhat spcnt part of the ycar in dens moved less Ihanothers, but non-denning classes of bcars did nOl differ int0l:11 annualmovelllent (AmstnJp 1995, Amstrup et al.2000).

Fcmales with cubs were generally lhe mOSI activegroup, ,lilt! single females the least active. Highest andlowest levels of ,lctivity were recorded in June andScptember, but there also was a strong activity peak inearly winter. Activity levels were lowesl in the earlymorning and higher from mid-day through late evening.

Beaufort Sea plllar bears kept their tIl0Velllelll.~ withinboundaries outside of which they seldom ventured.Annual activity areas ranged from 12,730 km~ to 596,800km1

• tlfonthly activity are:lS ranged from a mean of 344klll~ for females with cubs in April to 11,926 klll~ forfemales with yearlings in December (All1Slmp 1995,Amstmp et aJ. 2000).

Bears from the Beaufort Sea population occupied allarea extending up to 300 km offshore, from CapeBathurst in Canada to PI. Hope, Alaska, and enclosing939,153 klll1 (Amstrup et aJ. 1986, Garner et al. 1994,Amstmp 2(00).

Animals origin,tlly captured along the Beaufort Seacoast spent appro.>;illlately 25% ofthcir lime ill thellortllcastern Chukchi Se:l, but animals captured in lheChukchi Sea vcntured into the Beaufort Sca only 6% ofthe time. With few exceptions (Dumer and Amstmp 1995)bears captured in the Beaufort Sea were faithful toSllllllller activily areas in the central portion of theBeaufort Sea (Amslmp et 31. 1986, Amstmpl995,Amstmp et al. 1995, Amstrup et al. 2000). Although anybear caught in this region could be relocaled anywhereelse in the region, individual bears appeared faithful togencral geographic regions (Fig. 8.1). Rccelll analyses ofpallerns in scasonal fidelity of polar bears (Bethke et al.1996) suggested tllat 3 scparate populations or stockscould be distinguished.

Thesc 3 relatively discrete stocks overlap to a grcateror lesser extent within Alaska waters (S. C. Am.~trup, U.S.Geological Survey, unpublished data). Thercfore, it is nolonger re'l.~onable to refer to only I group of polar bears(Amstrup 1995, Amstrnp 2000) occupying this region(All1stmp et al. 2001). AltllOugh these groups arc notdistinguishable genetically (Paetkau cl al. 1999), thev arcdistinct enough to mandalc managllment rccogniti()J1~

Two groups, the Chukchi Sea and the SouthernBeaufort Sea populations, sharl~ the mainl;md CO;lSt,11areas of Alaska in the grllatest numbers (Amstrup ct al.2lX)l). Recognilion of these stocks helps to explain someof tlw lllo\"emem p<lllcIIlS previously ohservlld. Thesc 2gmups supply most of the harvest of pol;lr bears lhat

66 BIOLOGICAL SCIENCE REPORT USGSmRD 2002-0001

/./

Chukchin = 4977(125) ID®iii1..

90 r-T~l~ ,;/

~g := =P~T~ /~g --r~- ;~g40 -'~E-==, / 4030 ,-~~~ 35

18 ~"=-=- / ~go I / 20167D I

/./

//

/\

\\

Figure 8.1. Numbers and relocation positions of satellite radio-collared polar bears (# of individuals) captured in each of 6 longitudinalzones within the Beaufort Sea. Histograms illustrate proportions of those relocations made in each zone. For example, 32% of the 2,226relocations 01 bears originally captured in the Lonely zone were recorded in the Barter Island zone, Alaska; 47% of the 1,079 relocationsof bears captured in the Wainl'.Tight zone, Alaska. were recorded in the Chukchi zone.

OCCurs in Alaska and much of the harvest along themrlinland coast of northwestern Canada.

Oat;l were analyzed for .189 captures of 534 bearsbetween 1967-1974 (a period of hypothesized over­harvest) and for 1.087 captures of789 bears obtainedoctwcen 1981-1992 (:1 period when the population shouldhave recovered from over-ha!"vest). Population growththroughout the intervening years was also examined(Am5tmp 1995, Al11.~trup ct al. 200 I).

Amstnlp et al. CWOI) and t1cDonald and Amstrup(2001) suggested that the Humber of polar bears in theSouthern Bcaufort Sea popuilltion grew at more than )';';;,per year between 1967 and 1998. reaching an estimatedpopulation tll'lt could he as high as 2.500 animals.

Although contact with hydnw;lrbons Gin have scriOlISramifications for polar bears (Amstmp et al. 1(89). theJlolar bear's apparent rapid population growth hasspanned the entiro.~ history of petroleum de\'l'lopmcllt inarctic Al:lskil fAmstrup 2000, Amstmp et a1. 2001.i\lcDonald and Amstrup :WO\). This suggests thatmanaged resource development can be compatible withhealthy polar bear popUlations. Also ellcouraging h [he

lIew ability 10 estimate potential impacts that oil spillsmay have on polar bears. TIlnt ability has majorral1lificatiol1.~ for assessing risks of a variety of potentialdevelopments (Dumer et al. 200lb).

Both IOllg and short-term trends in condition ofindividual animals were observed during this study.Condition of adult females, as reOected by total mass.showed significant seasonal trends (Durner and AmstrIJp1996). Despite seasonal Ouctuations, longer-term trendsalso were suggested. Trends in recruitment and survivalrates (in the 1970s compared with those from 1980through 1992) suggested all inverse compensatoryrelationship hetween tlllal popUlation size and recruitmentof sllbadults. Popul:llion size alone explained ,SSt;:;, of thev.1riation in pmportions of 2- anti 3-year-olds in annualsamples (Armtrup 1995). Large populatiolls of the lallerpan of the slUdy appeared to recruit proportionately fewerju \'cniles, and snwller popUlations of the early pari of thestudy recntitetl higher proportiolls of juveniles.

Condition of single adult females and those willi cubs,;15 rctkctcd in mcasuremcnts of ;lxi;ll girth. appeared todecline significantly :IS the population grew. PI)jlulatio/J

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDUFE RESEARCH SUMt-.'fARIES 67

size alone explained 75% of the variation in axial girth ofreproductive age females.

Although numbers of young produced per femalewhen the population was small «0.40) and when it waslarge «0.38) were similar, Iillers of Illore than oneyearling were more frequell! when the population \Va.'ismall. Sampling inconsistencies during the 2 periodsprecluded comparison across years for cubs and 2-year­olds but not for yearlings. Observed reproductiveintervals of 3.4 and 3.7 years in early and late periodswere suggcstive of change, but not significantly different(Amstrup 1995). The :lge structure of the small population\\'a,'i younger than that of the larger population of lateryears.

Survival of aduHs, as calculated from life tables, washigher :lIId survival of yOLlng lower when the populationwas large. Survival rates of adult Beaufort Sea polarbears, however, were as high or higher than thosemeasured anywhere else. Annual survival of radio­collared females ranged from 0.946-0.980 (Amstrup andDumer 1995). Survival of cubs ranged between 0.610 and0.675, while that of yearlings was 0.751-0.903.

Intllis study hunting explained 85% of the ­documented deaths of adult female polar bears (Amstrupand DUnJer 1995). Natural mortalities were not cOllllllonlyobserved among prime age animals (Amstrup and Nielscn1989), and we still know lillie about Ille proxirn:lIc causesof natura} deaths among polar bears.

In the early 1990;;, the trends described abovesuggested a population that could be approaching carryingcap;lcity and was dtller swble or growing more slowlythan in the early 1980s. 1I.10re recent data suggest analternate hypothesis: ApP;lrent density dependence was afunction of more transitory ccological effects. Theapparent continued growth of the population illlo the latc19905 and the expansion of numbers of matcrnal dens aswell as expanded ureas ll.~ed for Jenning (st'(.' bdo\l')appear to contradict earlier conclusions regardingcarrying capacity and density effects. This suggests thatissues related to population statuS should be revisited(AllIstrup et a!. 1986, Arnstrup 1995, Am.,tmp et a!. 2001,tl-fcDonald and Alllstrup 2001).

Estill1:lIed numbers of bears at the close Ilf tile studywere relatively large. Effects of the increasing humanintrusions into the polar be;lr environr1lent lI;we not beenobserved at a population level, suggesting that pro;H.:tivemanagement can assure coexistence of polar bears ;lIldhuman developlllents.

AbsllllllC numbers of bears, howcvCl", still are ,mall<2ompared to many other .~pecies. Early estimatessuggested the additional loss of ;IS fev,; ·as 30 bears eachyear might push the total take from the population to them;lxillluill sustained yield (Amstrup et al 1986, Amstfup;md DeI\'faster 1988). Excess take did predpitate a declinein the 1%05 Hlld 1970s. H,:nce. although POPlJlatillJlS may

now be ncar historic highs, managers must be alert topossible changes in human activities, including huntingand habitat alterations that could precipitate futuredeclines.

Reproductive Significance of Maternity Denningon Land

The distribution of polar bears is circumpolar in theN{)fthern Hemisphere, bUlmatemal dens known at thestan of this project were concentrated in relatively few,widely scattered locations (Amstnlp 1986, Amstnlp et al.1986, Amstnlp and DdI'laster 1988, Amstrup and Gardner1994).

Among the best-known denning concentration areaswere the SVillbard Archipelago north of Norway; FranzJosef Land, Novaya Zemlya, and Wrangel Island inRussia; and the west coast of Hudson Bay in Canada.Denning W;JS cither unCOllllllon or unknown in gapsbetween known denning concentration areas. TheBeaufort Sea region of Alaska and Canada lay int11clarge.~t of those gaps, :md some had hypothesized thatpolar bears of this region actually were born in otherareas.

Now we realize that the coastal plain area of the ArcticNational Wildlife Refuge lies in a region of polar beardenning; :ll1d its coastal plain also may contain significantgas and oil resources. Polar bears in dens could beaffccted in many ways by petroleum development, butneither the distribution of dens nor the sensitivity of bearsin dens was known before our rcsearch (Al11stnJp 1986,Amstrup and DeMaster 1988).

To ;lscerlain the /lumber and distribution of Jenningpolar bears that could be impacted hy oil development onthe Arctic Refugc coastal plain, we cstablished thefollowing research objectives: I) to determine thedistribution of polar bear dens in northem AlaSKa, 2) toas<2ertain the time that polar bears enler and emerge frollldens, 3) to calculate the relative success rates of dens or]

land and 1)J1 sea icc, and 4) to determine whether oil andgas exploration and dcveloplllelll of the Arctic Refugecoastal plain would adversely impact polar bears of theBeaufort Sea by dismptillg denning activities.

Po!;lr bears were captured and radio-collared between1981 and 1992. Amstmp and Gardner (1994) determinedthat denning in the Beaufort Sea region was sufficient toaccount for the estil11;lIed pOJlulation. They :llso noted thatthe proportion of dens on land was. higher in the !nteI980s and early 1l)()Os than it was earlier in the study(Fig. S.2). 'nl;lt trend continucs, and other distributionalchanges al so Illay have occurred in the late 1990s andearly 2000s. Of a IOtal of IS2 dens located by telemetrybetween ,~pring of 1982 and spring of 2001, 150 werewithin the study area from 16]0 to 137" W longitude(Pointl-Jope to :'hJc.kenzie Rivcr). Polar bear dens in this

68 BIOLOGICAL SClENCE REPORT USGSfBRD 2002-0001

Figure 8.2. Number of polar bear dens localed by radio-lelemelry ineach of 3 substrates, 1981-1990.

region continued !O occur on land, pack iL:e, and land-fastice.

WCre II November and 5 April for land dellS and 22November and 26 lvtarch for pack-icc dens (Amstnlp andGardner 1994).

Female polar bears captured in the Bl:aufort ScaapPl:ared to be isolated from those caught cast of CapeBathurst in Canada. Bears followed to >1 den did nOIreUSl: siles, and consecutive dens were from 20 km to1,304 kill ap<lrt (Fig, 8.3). However, radio-collared bearswere u~'ually f:lithful to :;ub:;tr,llc (pack icc, land, land-faslice) and the general geographic area of previous dl:ns(Amslrup and Gardner 1994).

Of lhe 73 dens found by radio telemetry on themainland CO:lst of Ala.~ka and Canada (land plus faSI-icedl:ns), 32 (44%) were within lhe bounds of Ihe ArcticRefuge and 24 (33%) were within the 1002 Area.

The proponion of dens locatl:d on the Arctic Refugedropped frolll 47'70 10 41 % when lhe periods before :lndafter 1992 wcre compared, while the proponioll of denslocated within the bounds of the 1002 area dropped from36% 10 30%. The decrea.'>c in proportion of land dl:llS onthe Arctic Refuge was accomp3nied by an increase in theproportion of dellS found 011 land areas west of the ArcticRefuge. Although this diSlribution shift is 1I0t statisticallysignificant (Chi-square lest P = 0.88), it is readilyapp.:m:nt all the m.lp (Fig. 8.4),

The shift may be explained simply by sample sizelimitalions. The continuing growth in polar bear numbers,the continuing trend in proportion of dens 011 land, andperhaps changing freeze-up conditions in the hlst decadeall may be influencing the distribution of dl:nning efforts.111e apparent increase in numbcrs of bl:ars denning onland and the incrc,1scd land ,lren used for dl:nningcorroborates estimates, reponed earlier. that suggl:sll:d acontinul:d increase in tOlalnumbers of polar bears

DFASTlCEa LANDaPACKICE

14

~ 12

~ 10~

• e~.r; 6,8 4~

Seventy-three of the 150 mateOlal dens discovered bytelemetry between 16r Wand 137" W wefl~ on land orland-fast ice where they wcre potcntially vulnerable to avariety of human disturbances.

The remaining 77 identified maternal dens were ondrifting pack ice where they were relatively invulnerableto most hUlllan 3ctivities. 111e proportion of pack-icc densdropped dramalically in the lauer half of lhe study. Adecrease in slUdy elfort in offshore regions in the late1990s lIlay explain a portion of thc declinc inl\umbers ofdens found on p:lck ice. Bears denning on p<lck ice driftedas far as 997 krn while in dellS and were potentiallyvulnerable to n variety or natural forces thaI couldcompromise lhcir security while occupying dens(Amstrup and Gamer 1994).

There was no differencl: in cub production by be,usdenning on land and paL:k icc. I\.fcan enlry and exil dates

~21~~~20+.-,!Hi81 1982 1983 1984 19851986 19871988 1:1-891990

Ycar of DCI'll'lil'l9

,--

!--- ----1_,,

ji/

---1 .I

--I--,---

\

\-\

----- ---

Figure B.3. Maternal den locations for 5polar bears followed to dens for more than one ~ear. All dens were located by radio telemetry. Bearsrepeatedly denned inlhe same general geogmphic area, but not the same place. Likev,ise, polar bears repeatedly denned in the same substrate.(from Amstrup and Gardner 199.t)

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUt-.·ft-.1ARIES 69

170' 165' ,.,. ISS' ,,,. 145' 140' 135' ,,,., • ,

:- ~ , ", .,~, , ,, (', ,•, ,, .' , -$:

, '<\, . .,f/

.•~

\~, ,.., ,',

'",!< \ <8.

" '%i:-' ., ..~.. « , 0

;" .f, ,. ,. (,1 '1,·1] ", <t,,~

,-, .'. ,.0, ,. '. M. ,0", i. ,9-ff, ,

~ '" ,'. ,~

"" ('J,,.

8' ~,. ." ,%~,. ., ,.:2..g/ ~

, ,.,. ,i>

\~8' ,. ,., ,. "~, • '." ,.. ,

···Colville '":;;, .<, , , 0 ~, -, ,River .' McKenzie l e>Barrow <'~

/,, ,. ",

", Dclta ~ Rivcr , , ~

" ~), ,Odta

,, ,,,I 0 ,, ~<!~) ,, Point ,. , ,, ,

", Hope, '.'/\0 ,,. " •

0 , Maternity Polar Bear Dens ~,Discovered by Radio Telernetry,• Polar Boar Pen 1991 and earlier.--, Polar Bear Pen 1992 and later

IV Roads and Pipelines• 1002 Area of Ihe Arctic Refuge

165' 160' ISS' ISO' HS' 140'

FIgure 8.4. Distribution of maternal dens of radio-collared polar bears along the northern coast of Alas~a and Canada, 1981-2001,(updated from Amstrup and Gardner 1994)

throughout the sludy period. 111e distribution of lIlatemaldenning continues to be a fenile area for future research.

Despite a possible decline in proponionaluse of theArctic Refuge for denning, there still appears 10 be ahigher concentration of dens on the Arctic Refuge than onadjacent lands. DeveloplJlcnt of hydrocarbon resourcestllerdorc could incrC'\.~e the potential for disturbance ofdenning polar bears by human activities.

Because the chronology of denning is now known,howcver. human activities could bc temporally managedto minimize cxposure of denning bears (AmstOlp 1993,Amstmp and Gardner 1994). Spatialmanagelllcnt ofindustrial activities could further Illinimize exposure ofdens to disturbanc~sbecause Jenning occurs in lowdensity (including the Arctic Refuge) within relativelyuncomlllon habitats that can be mapped (Amstmp 1993,Dumer el al. 200Ia).

Available data indicate polar bears arc relativelyresilient to disturbance, coming from Ollt side their dens(Amstrup 1993. Amstrup ,lIld Gardner 1994). Datashowcd that dcns exposed to even high levels of activitydid nol suffer a delC'cwble reduction in productivity

(Amstmp 1993). Pcnurhatiolls resulting from capture,marking, and radio tracking nwtcmal bears did not affectlitter sizes or stature of eubs produced; and 10 of 12denned polar bcars exposed to exceptional levels ofactivity were not measurably affccted (Amstmp 1993).

Hcnce, polar bears in dens llIay be less vulnerable tohuman disturb,lnces than previously thought. TIlis findingcorroborates the oh.~crvadons or Blix and Lentfer (1992)who reported that polar bears in dellS arc well insulatedfrom disruptions outside of their dellS.

Aggressive and proactive management, therefore, canminimize or e1iminatc most of the potential adverseclfeets of hUlllan developmcnts on denning polar bears. Itwill be important to conduct rcsearch and monitoring ofpolar bear denning and ecology concurrent wilh anyapproved dcvelopments III assure that managelllent effortsdo have the desired mitigation effccl.~.

70 BIOLOGICAL SCIENCE REPORT USGS/BRO 2002-0001

References

Amstrup, S. C. 1986. Pobr bcar. Pages 790-80~ ill R.DiSilvestro, edilOr. Audubon wildlife report, 1986. TIteNational Audubon Society, New York, New York, USA.

__. 1993. Human disturbances of dcnning polar be;lrs inAlaska. Arctic 46(3):246-250.

__. 1995. Movcmcnts, distribution, and populationdynamics of polar bears in the Bcaufort Sea. Dissertation,Unj;'ersity of Alaska, Fairbanks, Alaska, USA.

__.2000. Polnr benr. Pages 133-157 in J. J. Truen and S. R.Johnson, editors. TIle natural hislOry of an af<;tic oi] field:development and biota. Academic Press, Inc., New York,New York, USA.

__, and D. P. DcMastcr. 1988. Pobr bear Ur.ws marilimlls.Pages 39-56 in J. W. Lentfcr, editor. Selected llw.rinemammals of Alaska: species accounts with research andmanngemcnt recommendations. "tarinc MammalCommission, Washington DC, USA.

__, and G. M. Dumer. 1995. Survival of radio-collared polarbears and their dependent young. Canadian Joumal ofZoology 73(7): 1312-1322.

__, and __, 1. Stirling, N. J. Lunn, and F. Messier. 2000.Movements and distribution of polar bears in the BeaufortSea. Canadian Joum<ll of Zoology 78:948-966.

__, and C. Gardner. 1994. Polar bear maternity denning inthe Beaufort Se<l. Joum:ll of Wildlife Management 58( I): 1­10.

__•__, K. C. Myers, and F. W. Oehme. 1989. Ethyleneglycol (antifree7.c) poisoning of a free·ranging polar bear.Veterinary and Human Toxicology 3 I:317-3 I9.

__, G. W. Gamer, and G. M. Dumer. 1995 Pobr bears inAlaska. Pages 351-353 ill Our living resources. U.S.Depanment of the Interior. National Biologkal Service,Washington DC, USA.

__, T. L. McDonald, and I. Stirling. 2(){JI. Polar bears in IheBeaufon Sea: a 30-year mark-recapture case histol)'. Jounlalof Agricultural, Biological, and Environmental Stathtlcs6(2):221-234.

__. and C. A. Nielsen. 1989. Acute gastric tlilatation andvolvulus in a free-living polar bear.lournal of WildlifeDiseases 25(4):601-604.

__, I. Stirling, antl J. W. Lcntfer. 1986.l'asl and preselllstatus ofpolnr bears in Alaska. Wildlife Societ)' Bulletin 1·1:241-254.

Bethke, H., 1\.1. K. Taylor, S. C. AmstnJp. and F. Messier. 1996.Population delineation of polar bears Ilsing satellite collar data.Ecological Applic<ltions 6( I ):311-317.

Blix, A. S .. and J. W. Lemfer. 1992. Noise and vibr;llion levels inartificial polar hear dens as relatet! to selected petro1cumexploration and deV(~lopmental activitlcs. Arctic ·15:20·24.

Dumer, G. M., ant! s. C. Amstmp. 1995. Movements of afemale polar be:lr frol11 thlrthern Alaska 10 Greenland. Arctic4X:338·3·11.

_~, :Jl\d _~. 1996. Mass and bolly.dimcilsionrelationships of polar bears in northern Alaska. WildlifeSociety Bulletin 24(3):480··t8.:1.

__•__. and K. J. Ambrosius. 2001,1. Remotcidentification of pol:Jr bcar malernal den habit;11 iJll1lmhernAlaska. Arctic 54(2): 115·121.

__•_~, and T. L. McDonald. 200lb. Estimating theimpacts of oil spills on polar bears. Arctic Research of theUnited States 14:33·37.

Gamer, G. w., S. C. Amstrup, I. Stirling, and S. E. Belikov.1994. Habit:ll considerations for polar bears in the NorthPacific Rim. Transactions of lhe North American Wildlifeand Nalllral Resources Confcrence 59: 111-120.

McDonald, T. 1.., and AmstnJp, S. C. 2001. Estimation ofpopulation size using opcn capture-recapture models.Journal of Agricultural, Biological, and EnvironmcmalStatistics 6(2):206·220.

Pactkau, D., S. C. AmstOlp, E. W. Born, W. Calvert, A. E.Derocher, G. W. Garner, F. Messier, I. Stirling, M. Taylor, O.Wiig, and C. Strobeck. 1999. Genctic structure of theworld's polar bear populations. 1..lotccular Ecology 8: 1571­15S.:I.

Stirling, I. 1990. Polar be;lrs and oil: ecologic perspectivcs.Pages 223-234 ill J. R. Geraci and D. J. SI. Aubin, editors.Sea Mamn1:lls and oil: confronting the risks. AcademicPress, San Diego, California. USA.

ARcnc REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH Sur...fMARIES 71

g• 30~

'"• 2000

"<I>-0 100"•~E,z 0

82 83 84 85 86 87 88 89 92 93Year

features snow gtese selecled when feeding (Hupp andRobertson 1998).

Figure 9.1. Numbers of lesser snow geese observed on the coastalplain of the Arctic National \'fIldlife Refuge, Alaska, USA. during aerialsurveys from 1982·1993. Poor weather prevented surveys in 1990 and1991.

Snow Goose Habitat, Food, and EnergyRequirements

TIle staging ;:Irea on the Beaufort Sea coastal plainprovides forage tll .. t geese use to build energy reservesprior to continuing their migration sOlllh (Pallerson 1974).Upon departure from the cO;lstal plain, snow geese make a2,000-klll-nonswp night to tlle next stopover area in110nhern Alberta (Johnson and I-lcner 1989). Geese thatlack sufficient fat reserves lllay be less likely to survivemigration (Dwell and Black 1991).

Because snow geese arc easily disturbed by humanaClivity (Davis and Wiseley 1974, Wiseley 1974, Belangerand Bedard 1989), development of the coastal plain coulddisplace geese frolll feeding habitats. Exclusion fromfeeding habit:lts could reduce the likelihood that staginggeese would acquire fat reserves needed for migration. Toidentify snow goose areas that could be impacted bydevelopmellt, we needed data on forage preference aswell as the distribution and avaihlbility of feedinghabitaL~.

We studied body condition and diet of snow geese inorder 10 understand their energetic and nutritionaldemands. Wc also ~lssessed usc and availability of fcedinghabitats and the effects that grazing geese h;ld onvcgetation at the.~e sites.

Body condition and diet tlf l.'i I snow gcc~e collct:teddurin!:! 1984-85 ;UH119K8 were cvalll;lted (Fig. 9.3). Adultsl\ow-gecsc gained an average of 22 g of body f;lt/day anddeparted the Arctic Refuge with about 600 g of falreserves. Juveniles ;lrrived on the Art:tic Refuge withsrna!lt:r fat reserves than adults, acquired lipid reserves ata slo\\"er rate (I] g/day), and departed with smallerreserves (375 g). At the end of staging, juvcniles lj;\d

Section 9: Snow Geese

Part of the coastal plain of the Arctic National WildlifeRefuge, Alaska, is used as all autumn staging area bylesser snow geese (Chell cacru{c.fcens c(/nllh'.fc~'!ls) fromthe Western Can;:ldian Arctic population (hereafter calledthe Westem Arctic population). There were approximately200,000 breeding adults in the Western Arctic populationthrough the mid-1980s (Johnson ~llId Herter 1989), btlt thepopulation has recently increased to about 500,000breeding adults (Kerbes et a!. 1999).

Early in their autulllllllligration, adult and juvenilesnow geese from the Western Arctic population feedintensively while staging on the Beaufort Sea coastalplain in Canada and Alaska to build fat reserves lJ(;:ededfor migr~ltioJl. Aerial census!!.s from 1973 to 1985indicated that up to 600,000 adult and juvenile snowgeese used the coastal plain for 2-4 weeks in late Augustuntil mid-September (Oates et al. 1987).

We studied arlnual variation in numbers and sp;lti,lldistribution of snow geese tlwt staged on the coastal plainof the Arctic Refu£e.

Numbers and distribution of SIIOW £eese on the ArcticRefuge were assessed from aerial surveys during 9 yearsfrom 1982-1993 (Robertson ct al. 1997). During surveysbiologists estimated the numbers of geese in flocks andmarked flock locations olllOpographic maps. Surveyresults were digitized on a lllilJl of the coast:d plain. A gridof 25-km" cells was superimposed over the digitized map.We t~llJied the numbers of geese observed in a cell duringeach survey and the /lumber of years each cell W~lS uscdby geese.

The numbers of snow geese that staged on the ArcticRefuge ranged frolll 12,800 to 309,200 individuals acrossyears (Fig. 9.1). The numbers were highly variablebecause in some years most of the population remained inCanada. whereas in other years the majority of theWestem Arctic population staged on the .·\rctic Refuge.

Snow geese occupied approximately 605,000 ha of theArctic Refuge coastal plain between the I-Iulahula Riverand Canadian horder. Only 20',:;" of the 25-klll l t:ells wcrefrequently u.~ed (i.e., used ~j years) yet 80% of thefrequently 1I,eu cclb fell within the boulldarie.~ of theJO02 Area (Fig. 9.2). The mid·coastal p!;lin,!letween thcOkpilak and Aichilik rivers was used more frequentlythan areas ncar the coast or the steep foothills. ArC;lS thatwere used frequently were :l!sO llsed by larger Illllllbers ofgeese. Frequently used ;lreas had Illore of the landscape

Size and Distribution of Snow GoosePopulations

Jerry n~ lillpp, OOlllla G. Robnt.'lo/l, am! A/all IV.Brackney

72 BIOLOGICAL SCIENCE REPORT USGSffiRD 2002-0001

_ High Use,AreI1 (5 or more }'\'anI)_ Medium Use Area (3 to -I years)_ Low Usc Area (1 to 2 yell1!:i)r:::J An.1.k NWR l..and.5rZJ Kaktovik Jnuplat Corp. u.nds- - 1002 Boundary

A

-_.:::::", ------=--Lesser Snow Geese

Use Areas, 1982 - 1993o S 10 15 20m!

IJ 8 J6 U ~km

N

FIgure 9.2. Frequency 01 use of 25·km1 cells by lesser snow goose flocks on the coastal plain of thQ Arctic National 'Nildlife Rduge, Alaska,1982-1993. Use of ceils by snow geese was assessed during aerial SUlveys.

proportionally smaller lipid resen'es (15-18% of bodyIllass) lhan adults (21-24':',;" of body mass) and likely wereal greater ri.~k of having inadequate energy reservcs formigration.

We examined esophageal contenls of snow geese toidelltify important forage species (Brackney and Bupp1993). Snow geese primarily consumed 2 food items: theunderground stcmbase of Eriophorllm IJIlgusrijolillll1 (tallcouongrass) and the aerial shoots of Eqllist'1U1IIvariCglll!l1Il (northern scouring-rush). The birds typicallyfed on northern scouring-nlsh during the morning whellsurface soils and Wall:[ were frozen, and they consumedunderground parts of tall coUongrass during afternoonand evening ancr soils had thawed.

We examined forage illlake and digestibility amongcaptive snow geese 10 beller understantlthe population'sforage requirements (Hupp et a!. 19(6). Snow geese fedfor a high percentage of the day (5{)'(iO';""r) and maintainedhigh rates nfforage intake (14 g dry matter/hour). On adaily hasis a goose probably consumes the equivalellt ofabout 3Wi, of its body llIass in clltlOllgrass stl'lllbases, :\

population of 300,000 snow geese that stages for 3 weekscould consuillc as much as 4,200,000 kg (wet mass) ofcO!longrass stembases. Thus the population consumes avery large amount of forage in a short period.

Nonhem scouring-rush primarily grew on riparianterraces within 400 m of river channels. Riparian terracesadjacent to rivers arc important habitat for snow geese asIlley feed on scouring-msh.

We measured vegetation and soil moisture at siteswhere snow geese fed 1m lall cOllongrass, and then wedevdoped a statisticall1lodelto identify suitable feedinghabitat (llujlp and Robertson 1998). Thc model was testedusing captive snow geese. Snow geese typically e.\ploitedsmall, homogeneous patches of cOlltlngrass in floodedareas. 'Illey avoided uplands and flooded areas wherecOllongrass was imermixed with Cafex, shrubs, ortussocks.

The habit'll selection model was used to assess theavailability of couongrass feeding sites along 192randomly located transects on the 1002 Area cast of thelIulahula River. Cottongrass fceding sites occurred in

ARCTIC REFUGE COASTAL PLAIN TERRESTRIAL WILDLIFE RESEARCH SUl....IMARIES 73

ADU.TS."

§6Xl •III ·i .00: "'" •:J .. 0

20 Y=159.1 +35.3x - 0.65:<

o FEI>WES ."""'"0 5 10 15 '" 25

reduces forage availability in subsequcnt years. Geesemay be unable to successfully exploit a site for severalyears after it has been grazed.

Snow geese consume large volumes of forage atfeeding sites that are slll<lll, patchy, and comprise .:53% ofthe l:l.lldsc<lpe. Feeding on cotlongrass at a site reducesforage abundance at that location for al least severalyears. Snow geese in the Western Arctic population usc anextensive staging area because forage availability variesboth spalially and temporally. Variation in the numbers ofstaging geese on the Arctic Refuge is likely due to annualdifferences in habitat conditions. Poor forage conditionsor the presence of snowcovcr on the Canadian portion ofthe staging area lIlay contribute to grealer number ofstaging geese on the Arctic Refuge.

FIgure 9.3. Rates 01 lipid deposition by lesser snow geese during fallstaging on the Arctic National Wildlife Refuge, Alaska. Geese werecollected in 1984, 1985, and 1988. Data were pooled across years andsize of fat reserves scaled to the date geese were first obselVed on theArctic Refuge in each year,

Effect and Mitigation of Human Activities onSnow Geese

Staging snow geese are easily disturbed by aircraftactivity (Davis and Wiseley 1974, Belanger and Bedard1989). Repeated aircraft disturbance can reduce their rateof food intake due to disruption of feeding behavior anddisplacement from feeding habitats. Reduced fataccumulation and diminished survival during migr3tioncould result from repeated aircraft disturbance.

The following objectives were designed to llssesssnow goose response to experimental aircraft overflights:I) determine the effect of aircraft on activity patterns andhabitat usc, 2) calculate the effect of increased stress ordisplacement caused by aircraft overflights on the energybudgels of the geese, and 3) determine implications ofpetroleum devcl(lpmelll to survival of snow geese.

Studies of aircraft disturbance were limited due to lownumbers of geese on the Arctic Refuge in most ycars from1988-1993. poor weather, and the need to meet otherstudy objectives. Snow geese flushed at a mean disl;lnceof 5.2 km (SD = 2.Y) from a Bell 206B helicopter duringoverl1ights in 1991 (11= 19). Flocks were displaced anaverage of 1.8 km (SD = 2.0) froilltheir feeding sites.

TIlese results arc similar to a 1973 study of aircraftdistmhance to the Western Arctic population in Canada inwhich fixed-wing aircraft and helicopters l1ushed snowgoose Hocks wilhin a 6-km radius (Davis and Wiseley1974). In that study, l10cks were disphu.:cd an average of1.9, 1.6. and 5.9 kill frolll feeding sites by helicopters,small, and large fixed-wing aircraft, respectively.

Several studies suggest that human disturbance candisplace staging snuw gee.~e from feeding habitats andpossibly diminish the size ofjuvcnile fat reserves. A studyof staging greater snow geese found that >2 disturbances!hom caused 50% fewer geese to usc the disturbed area!11e following day (BClanger and Bedard lYH9). Energeticreserves of juvenile snow geese staging on the coast ofthe Beaufort Sea in Canada were projecled to diminish

2510 15 20

DAY OF STAGING5o

." JlNENILES

:§

III."

'" •0: y= 101.5 +13.1x • ••:J "'"0

""

snHlll patches that were highly interspersed with lesssuitable feeding habitat. 'Il\l~Y were widely distributed butcomprised a small percentage (.:53'}i;,) of the study area.

Larger-scale micro-relief features weft~ also examinedat SIlOW goose fceding areas. COHongrass feeding sitesprimarily occurred along Ilarrow « I 1lI) edges of Hoodedthermokarst pits, water tracks, and troughs. When feedingon cottongrass, snow geesc selected area.~ with greateravailability of therlllok<!rst pits and avoided uplands, lowcenter polygons, wet meadows, and strangmoor (Huppand Robertson 1998).

'nlernwkarst pits and water tracks were widelyavailable in tile mid-coastal plain between the Okpilakand the Aichilik rivers. Greater availability of co[[ongrassfeeding habitat in that region likely accounts for its morefrequent usc by snow geese (Bupp and R9bertson 199:\).

Snow geese removed the undergroiintl portion ofcotlongrass frolll which plants rcgenerate. Four years afteran expaimemal removal, the biolll<lsS of stembases intreatment plots was approximately 50% of that in controlplots (Hupp et al. 2000). Feeding by snow geese likely

REPORT DOCUMENTATION PAGE FOllT1l1pprovedOMB No. 0704~168

Public rtportlng burden for ifill colltetlon JI nUmaltd to Ivtrlll11 hour ptr ,"pon.., Including time for reviewlrJil imlructlons, searching e.dstlnlldata sources,g~ and malnlalnllv the data oeeded. and COOlple~r"9 and revlewlng the collection of Informallon. ~nd comments regardinll this burdene1linate or any othet aspect of this Mecllon of 1nl00000lon,lnduding Sl1J9llSllom fOf rllducing this boo:!en, to Washlr9lon Headquartm Ser..lces,Directorate for Inlormatlon Operations and Reporb, 1215 Jefferson Davis Hl;lhway, Suile 1204, Allington. VA 22202""'302. and to lhe OfrK:e of Managemenland Budge!, PapefWOl1( RedvcUon Projoc:t j0704..()168) WaslllrYJlon. DC 20503.

1. AGENCY USE ONLY (laM BIn) 2. REPORT OATE 3. REPORT TYPE AND OATES COVEREDApril 2002 Biological Science Report

4. TITLE AHD SUBTITLE 5. FUNDING NUMBERS

Arctic Refuge CO:lStal Plam Terrestrial Wildlife Research Summaries

6. AUTHOR(SI

Douglas, D. C, RC}Tlolds, P. E., and Rhode, E. B" editors

7. PERFORMING ORGANIZATION HANE(S) AHO AODRESS{ES) a. PERFORMINGORGANIZATION REPORT

U.S. Department of the Interior NUMBERU.S. Geological Survey

USGSiBRDiBSR·2002-0001

9. SPOHSORrNGfMONITORlNG AGENCY NAME(S) AND ADDRESS{ESj 10. SPOHSORIHG.MONITORINGAGENCY REPORT HUMBER

U.S. Department of the InteriorU.S. Geological SUr\'ey

11. SUPPLEMENTARY NOTES

12.1. DISTRIBUTiON/AVAILABILITY STATEMENT 12B. DISTRIBUTiON CODERelease unlimited. Available from the National Technical Information Service, 5285 Port ROy:llRO:ld, Springfield, VA 22161 (1-800.553-6847 or 703-487-4650). AVlIilable to registered usersfrom the DefenseTechnicallnformation Center, Attn: Help Desk, 8722 Kingman RO:ld, Suile0944, Fort Belvoir, VA 22060-62 I8 (1-800-225·3842 or 703-767-9050).

13. ABSTRACT (Mal:m\JTl 200 WOrt!!)

This report summ:lw-es investig:llions ofkcy terrestrial wildlife spl.:cies that use an appro;<;imate 1.5-million·acre al"C<l ofthe ArcticNational Wildlife Refuge costnl plain, c:llIcd the" I002 Area," in northeastern Alaska. This report contains information aboutpopulation dynamics, diSlnbution, energetics, and habitat use of the 1002 Area by canbou, musko;<;en, polar bears, and snow geeseas well as gri7.zly bears, wolves, and golden eagles. The report includes discussions of the potenlial effects and mitigation ofpetroleum explomlion or development for these species and lheir habitats.

14. SUBJECT TERMS (Ke)WOl'ds) 15. NUMBER OF PAGESArctic National Wildlife Refuge, ANWR, 1002 Area, Alaska, North Slope, tundra, ix+ 75coastal plain, caribou, Porcupine caribou herd, muskmen, polar bear, snow geese,

16. PRICE CODEbrown bear, petroleum

17. SECURITY CLASSIFICATION 11. SECURITY CLASSIFICATIOH 19. SECURITY CLASSIFICATIOH 20. LIMITATION OF ABSTRACTOF REPORT OF THIS PAGE OF ABSTRACTUnclassified Unclassified Unclilssified Unclassified,

Technical Report Seriesu.s. Geological Survey

TIle Biological Resources Division publishes scientific and technical articles and reports resulting from the researchperfonncd by our scientists and partners. These articles appear in professional joumals around the world. Reports are

published in two report series: Biological Science Reports and Infonnation and Technology Reports.

Series Descriptions

Biological Science Reports ISSN I081·292XThe series records the significant findings resulting fromsponsored and co-sponsored research programs. They mayinclude extensive data or theoretical analyses. Papers in thisseries nre held to the same pccr·rcview and high qualitystandards as their joumal counterparts.

Informalioll ami Technology Reports ISSN 1081-29111l1~s~ reports are intcnded for publication of book-lengthmonographs; synthesis documents; compilations ofconference and workshop papers; important planning andreference materials such as strategic plans, standard operatingprocedures, protocols, handbooks, and manuals; and datacompilations such as tables and bibliographies. Papers inIhis series arc held to the same peer-review and high qualitystandards as their journal counterparts.

U.S. Geological Survey· Alaska Science Center1011 E. Tudor Road, MS 701

Anchorage, Alaska 99503

ChierWilliam K. Seitz

Production Staff

Document Layout andCovcr Dcsign

Section I Map Figurcs

Digital DistributionAssistancc

Cover Photography

Mary E. Whalen

Jay A. Johnson

James E. I-Iaga and Adeline C. Sheridan

Coastal plain: A. Thayer; caribou: K. Whitten;muskoxen: M. Elllers; grizzly bcar: H. RC)110Ids;snow geese: 1. Bupp; \'cgctatioll sampling: C.Curby; polar bears: S. Amstrup; tussockcOtlongrass: M. Elllcrs.