Document yrZQVQg7aG5yerkx0ZE9k0d6V

ARCTIC REFUGE COASTALfPLAIN TERRESTRIALfWILDLIFE RESEARCH fUMMARIESf 31f single hypothetical development scenario presented in thef 1987 Final Legislative Environmental Impact ftatementf (Clough et al. 1987). The scenarios in Tussing find Haleyf (1999) are based on the most recent estimates of thef distribution and quantity of oil reserves within the 1002f Area (U.S. Geological Survey 2001).f This protocol assumed oil field design similar to thef Kuparuk and Milne Point petroleum development areasf within the scenario boundaries. The modeling exercisef could be Used fo fissess fhe potential effects 6f fidditionalf development scenarios that are not presented in Tussingf and Haley (1999) or Clough et al. (1987).f Central Af rctic herd parturient females actuallyf separated their concentrated calving areas fromf development infrastructure by about 7-8 km (Wolfef 2000). Wefused fi conservative displacement bf Efkmf based fin observations by Cameron fit fil. (1992) 6f increased caribou density from 4 km outward beyondf roads and pipelines. Calving sites fand the entire annualf calving grounds were displaced along with thef concentrated calving areas.f Our protocol stated that a concentrated calving areaf could not be moved onto fhe Beaufort Sea. Wefmade nof changes in shape of fhe concentrated calving areas orf annual fialving grounds. JAs fi fesult of fhese shifts,f relatively small portions of fhe peripheral, fow-densityf calving areas were occasionally moved onto the Beaufortf Sea along with some associated calving sites. Wf eftreatedf these ocean sites as missing data when assessing thef potential effects of displacement on fialf fiurvival.f Modeled displacement for fhe Porcupine fiaribou herdf was fo the Cast find fiouth, parallel fo fhe Beaufort feaf coastline, because that fs fhe direction of fhe herd'sf migratory approach to the fannual calving grounds inf spring. fDisplacement of fhe developed-zone fioncentratedf calving areas of the Central .Arctic herd has beenf primarily to the south, the direction of approach to thatf calving ground from winter range.f Our protocol minimized displacement of thef Porcupine fiaribou herd fialving grounds into fhe foothillsf and mountain zone. Tf his ftended to keep the annualf calving grounds on the coastal plain in the best remaining foraging habitats. fn fiome fiases, observed fioncentratedf calving fireas (e.g., fn f988, 2000, find 2001) did notf overlap the boundaries of finy of the hypothetical development scenarios, find fn fhose fiases the finnualf calving ground was not displaced.f Once the concentrated calving fireas and associatedf annual calving grounds and calving sites were displaced,f the forage during peak factation (NDVI_621) within fhef displaced annual calving ground was re-inventoried, thef median was recalculated, and fhe proportion of calvesf born in the low predation frisk zone (coastal plain) wasf recalculated.f Figure 3.28. Estimated change in calf survival during June for fhez Porcupine caribou herd, 1985-2001, as a function of the distance ofz displacement of the annual calving ground and associatedz concentrated calving area and calving sites. Upper and lower dashedz lines indicate 95% confidence intervals on the mean effect.z Then the empirical model was again used to predictf calf survival for the displaced calving ground. Thef difference between fhe fialf survival estimate for fhef displaced and observed calving ground was calculatedf and fi dataset of 46 displacement distances find fissociatedf changes in calf survival was generated for analysis.f The model showed a significant (r2 = 0.47, PbQ 0.001) inverse relationship between displacement distance andf predicted change in calf survival (Fig. 3.28).f The simulations indicated that a substantial reductionf in fialf survival during June would be expected finder fullf development of the 1002 .Area. Eighty-two percent of observed calving distributions would have been displacedf and the average distance of these displacements wouldf have been 63 km (range 16-99 km). This would havef yielded a net average ef ective displacement of 52 km andf an expected mean reduction fn fialf survival of 8.2% (SEf i fi.7%).f It fs remotely conceivable that fialving fiaribou of fhef Porcupine caribou herd could fselect habitats that yieldedf equivalent forage and predation risk after displacement.f Forage for factating females of fhe Central .Arctic herd,f however, fdeclined as the concentrated calving area in thef developed Zone shifted fo the south-southwest (Wolfef 2000). This suggests fhat such compensatory habitat usef by the Porcupine caribou herd would be unlikely if theirf calving fgrounds fwere fdisplaced fby foil fdevelopment.f Because there was no empirical basis for changing thef shape of the observed fialving distributions, it wasf impossible to estimate the magnitude of the ef ect of considering the peripheral calving fareas and calving sitesf as missing data when they were displaced onto the ocean.f The ef ect was expected tfo be small. Af rbitrarily assigningf calving sites that were displaced onto the ocean back ontof the coastal plain and making no other adjustments wouldf 32f BIOLOGICALfSCIENCE REPORTfUSGS/BRD 2002-0001f have increased displaced calf survival by only about 0.6%f on fiverage. This probably Constituted fhe fnaximumf possible effect of freating fireas find Calving sites fhatf were displaced to the Beaufort Sea as missing data.f tochastic simulation modeling (Walsh et al. 1995) indicated that fi 4.6% reduction fn Porcupine caribou herdf calf survival during June, all filse held equal, would havef been sufficient fo halt growth bf fhe Porcupine caribouf herd during the best conditions observed to date. Af10-kmf average displacement in our simulations would have beenf sufficient fo bring fhe upper confidence Interval bn fhef mean effect below a 0% predicted change in calf survivalf (Fig. 3.28). Afnean displacement of 27 km in ourf modeled predictions Would have been sufficient fo feachf the threshold of 4.6% mean reduction fn calf survivalf sufficient fo halt growth ff fhe Porcupine caribou herdf under best observed growth conditions fo date. This latterf level of displacement could occur well before fullf development of the f002 Area.f The estimated effect 5f displacement df fhe Porcupinef caribou herd 6n calf survival during June wasf conservative for several reasons. First, we used fhef conservative estimate of a 4 km displacement of concentrated calving areas fromfinfrastructure (Cameronf et fil. 1992) versus 7-8 km (Wolfe 2000). Second, wef displaced the concentrated calving areas parallel to thef Beaufort Sea coastline thus maintaining calvingf distributions on the best remaining coastal plain habitatf and minimizing displacement into fhe foothills wheref predation would be expected to increase calf mortality.f Finally, frelatively lfow fdensity fcalving wf as afllowed tfof overlap fdeveloped fareas, fas fhas fbeen fobserved for tfhef adjacent Central .Arctic herd (Wolfe 2000, Lawhead findf Prichard 001).f Because the assumptions were conservative, thef results were conservative. Substantial (10 to 27 km) displacement of concentrated calving areas and associatedf annual calving grounds and calving sites of the Porcupinef caribou herd fs fikely fo negatively fiffect calf survivalf during June. At fhe upper end 6f this range 6f displacement (27 km), recovery of the herd from thef current decline (Fig. 3.8) would be unlikely.fThesef conclusions fire consistent with fhose found fn fhe f987f Final Legislative Environmental fmpact ftatementf (Clough et al. 1987).f The Porcupine caribou herd has demonstratedf substantial natural variability fn size find demographyf (Figs. 3.5, 3.8, 3.10a-c). Because development 6f fhef 1002 .Area would fake fime, finy effects fin fhe herd'sf performance may take decades to detect. Reduced calf survival may slow the rateM increase during positivef phases 6f the growth curve 6f the herd find increase fhef rate 6f decline during fhe negative phases 6f fhe herd'sf growth curve. The period of natural cycles in herd sizef may increase and the amplitude of herd size may bef af ected.f The best empirical tool available for detectingf potential effects 6f development fs fhe fnodeledf relationship between calf survival find forage for femalesf during peak lactation demand (NDVI_621) within thef extent of calving (Fig. 3.26). This model is independentf of fictual finnual calving ground focation find fincompassesf a near full cycle of herd size as well as substantialf variation in hemispheric weather patterns (Fig. 3.5) andf variation in calving ground location (Fig. 3.13).f With industrial development, if observed calf survival falls below fhe fower 95% confidence Emit fin fhef predicted observations from this model (Fig. 3.26), or if af parallel pattern of calf survival yields a significantlyf lower Intercept ferm, fhen fin ef ect of development onf calf survival would be indicated.f Individual observations that fall below the lowerf confidence limit and which fcan be satisfactorily explainedf by exceptional environmental characteristics fe.g., carry-f over effects 6f near-catastrophic conditions in 1992 tof 1993 fifter eruption 6f Mount Pinatubo) (Fig. 3.26) needf not be considered evidence for effects 6f development onf calf survival. AJpattern of observed calf survival belowf the fower confidence Emit would be cause for concern.f tatistical methods for making these types of decisionsf are currently in development (Rexstad and Debevecf 2001). This assessment will require continued fntensivef calving ground surveys find calf survival dstimates.f Conclusionsy Our research has shown that fhe Porcupine caribouf herd has significant annual variance in calving groundf location (Fig. 3.13), faces finnual variance in habitatf conditions, selects areas with abundant high quality forage for calving, has Increased survival 6f calves bornf in the concentrated calving areas, and shows a correlationf between calf survival find both forage for females duringf peak lactation and predation risk in the annual calvingf grounds. Af ll this implies that unrestricted access to annualf calving grounds and concentrated calving areasf maximized performance of lactating Porcupine caribouf herd females find fheir calves. Because fhe Porcupinef caribou herd has shown limited capacity for growth, freef access tfo fcalving fground fhabitats mf ay fhave fcompensated for less than optimal wintering habitats.f Location of the concentrated calving areas during thef past 19 years (1983-2001) is the best estimate of the areaf that has provided the fhighest quality calving habitat for females and their calves. Calf survival within thef aggregate extent of concentrated calving areas has beenf higher than for calves born ifn areas never used as af concentrated calving area (83.8% vs.73.9%, respectively,f