Document 15q4p2V4oD6ZDnr7w2MV6EarE

ARCTIC REFUGE COASTALfPLAIN TERRESTRIALfWILDLIFE RESEARCH fUMMARIESf 27f the coastal plain with lower suspected density of wolves,f eagles and bears; and greater (8.3%, Pb= 0.026) if born inf the 1002 Area.f The survival advantage of high density Calving fof individual calves tended to be greater when calves weref born in the foothills and mountains than when they weref born on the coastal plain (14.3% advantage vs. 7.9%f advantage, rfespectively).f Individual calf survival was not related (P = 0.160) tof the frequency of use hf Its birth Site s a portion hf thef concentrated calving area, 1983-1994, but calf survivalf was lower (9.9%, Pb= 0.026) if the birth site was in anf area never used as a concentrated calving area. In af stepwise fogistic regression analysis fhat fimultaneouslyf considered calving density, time of birth, zone of birthf (coastal plain hr foothills), and fn or out bf fhe 1002 Area,f only calving density fP=f).004), fime period of birthf (early, middle, fate; PZ=10.012), and Zone fP=f0.008) entered the model that predicted individual calf survival,f 1983-1994.f The Survival advantage of both high calving densityf and being born near the middle of the calving period mayf have been due to predator swamping where high spatialf and temporal densities of calves may make it dif icult forf predators to capture individual calves (Hamilton 1971).f Bears fended fo be fess successful fit capturing calves fnf the concentrated calving areas of the Porcupine caribouf herd (Young find McCabe 1997).f When assessing the proportion of the annualf population of calves that survived during June, the timingf of birth in relation to other calves was not applicable, butf median calving date, 1983-1996, was available. Inf addition, we could consider the felative amount of foodf (NDVI_calving, NDVI_rate, find NDVI_621), winterf range conditions prior fo calf birth (snow properties), findf the proportion of calves born fn coastal plain or foothillf zones.f Analyses of the proportion of calves surviving inf relation fo fhese fndependent variables were conductedf separately at 2 scales : a) the extent of calving and b) thef annual calving grounds.f Within the extent of calving, the relative amount of orage available to females during peak lactationf (NDVI_621) provided the best model of calf survivalf during June (r2 = 0.85, PM0.001) (Fig. 3.26). No otherf independent variable that was considered addedf significant explanatory power.f This model (Fig 3.26) (Percent June Calf Survival =f [0.107 > (2.05 R NDVI_621 fn the fixtent of calving)] Rf 100) was the best available estimate of survival of calvesf during June for the Porcupine caribou herd finderf undisturbed conditions fluring the past 2 decades. Thisf model of calf survival was fndependent of annual calvingf ground focation and, ff the 1002 .Area fs developed, fhef Figure 3.26. Calf survival through June for the Porcupine caribouz herd, 1985-2001, in relation to median Normalized Differencez Vegetation Index on 21 June (NDVI_621) within the aggregate extentz of calving (EC). Legends identify the year of the estimate. Calf survivalz was not estimated in 1986 because inclement weather prevented az complete sample in late June. Calf survival for 1993 was a significantz outlier (RStudent = 3.84, see text for biological justification) and wasz xcluded from the estimated regression line (r2 = 0.85, P < 0.0001).z Upper and lower dashed lines indicate 95% confidence intervals on thez predicted observations.z model can be used to assess whether calf survival duringf June fs fiffected by development.f Calf survival for 1993 was an outlier (RStudent =f 3.84) and excluded from fhe estimated relationship! between NDVI_621 fn fhe extent of calving and calf survival (Fig. 3.26) and from fill subsequent models of calf survival. During 1992, atmospheric fierosols from fhef eruption of Mt. Pinatubo fn fhe Philippines reached fhef Arctic in the spring (Sfone et al. 1993). This resulted in af late spring, cool summer,fearly and heavy fnowf deposition fn fhe fall, find near catastrophic conditions forf caribou.f Wefsurmise that the consistently bad weatherf conditions during 1992 and early 1993 resulted in a carry-f over ef ect that reduced calf survival in 1993 to levelsf much lower than would have fbeen expected on the basisf of NDVI_621 alone. ft was fikely fhat fhis suspectedf additional mortality in 1993 af ected calves within the first day or two of life; perhaps many calves were of veryf low birth weight. Wefdraw fhis conclusion because 0- tof 3-week weight-gain of calves fhat survived fo be radio-f collared in 1993 was as high as any other year (Fig. 3.23) and the weights of parturient females that were caughtf with their live calves on ~21 June fn 1993 were as high asf any weights we observed, 1992-1994 (Fig. 3.25).f At the smaller scale of the annual calving grounds, thef proportion of Porcupine caribou herd calves that survivedf through June was positively related fo both NDVI_621 fnf the annual calving grounds and to the proportion of calvesf that were born on the fcoastal plain (assumed lowerf 28f BIOLOGICALfSCIENCE REPORTfUSGS/BRD 2002-0001f predation risk) (r2 = 0.70, PbQ 0.001). No other variablef predation on calf survival, it fs imperative to specify thef added Significant explanatory power. Median NDVI_621f scale of analysis, and assess multiple scalesf in the annual calving grounds and the proportion of calvesf simultaneously.f born on the coastal plain were not correlated (Pb 0.94).f The temporal increase in forage during peak lactationf Forage fn the annual calving ground accounted forf (NDVI_621) (Fig. 3.4) was coincident with local climatef approximately 75% of the total variance explained by thisf warming (Fig. 3.3 a). Forage at calving (NDVI_calving) model and assumed predation fisk accounted for fhef was positively associated with fhe .Arctic Oscillation (Fig.f remainder (Fig. 3.27).f 3.6).fThere were also positive relationships betweenf Thus, fn addition fo Scale dependency fn fhe functionalf climate and NDVI_calving, between percent o femalesf response hf caribou fo habitats (selection Cf fNDVIsf calving in the 1002 .Area and NDVI_calving, and betweenf within the extent of calving and within the annual calvingf calf survival and NDVI_calving [r2 = 0.33, Pb 0.011f grounds), there was scale dependency in the numericalf (annual calving ground); r2 = 0.60, PbQ 0.001 (extent of response of calf survival fo calving ground location andf calving)]. As a result, June calf survival was weaklyf habitat conditions. Only forage was related to calf correlated (r2 = 0.22, Pb=f).029) with fhe proportion of survival at fhe fargest spatial scale (extent of Salving) fhatf cows fhat calved fn fhe 1002 .Area. Further, becausef we analyzed.f climate af ected calving ground location (e.g., Porcupinef At the intermediate scale (fannual calving ground), caribou herd females were more likely to use the westernf forage dominated calf survival, but predation risk addedf portion of the extent of fcalving following winters with af substantial explanatory power.fAt the smallest scalef positive Af rctic Oscillation), both forage availability andf (individuals within the population of calves), spatial andf predation fisk were implicitly related fo flimate.f temporal variance in calf density (indirect predation risk) In years with substantial snowcover on the coastalf and direct predation risk most ef ectively explained calf plain (Fig. 3.18) and relatively low NDVI_621 in thef survival.f extent of calving, average calf survival (66%, Cb= 7, SE =f This scale dependency in calf survival likely occurredf 6%) was 19% less (Pb= 0.008) than when there was littlef because the annual variance in habitat conditions in bothf snowcover at calving and NDVI_621 was high f85%, Cb=f the extent of calving and in tfhe annual calving grounds farf 6,f Ef=f11%). Thus, climate was an important influencef exceeded the annual variance in predation risk within thef on habitat conditions, on the likely fuse of the Af laskaf extent of calving and within the annual calving grounds.f coastal plain and 1002 .Area for calving, and Sn calf The scale dependency in calf survival made it impossiblef survival during June, 1983-2001, finder undisturbedf to extrapolate across scales. Thus, fo develop anf conditions.f understanding of the relative influence Cf forage andf Potential Effects pf Development on June Calfy Survivaly Figure 3.27. Predicted calf survival for the Porcupine caribou herd,z 1985-2001, in relation to median Normalized Difference Vzgetationz Index on 21 June (NDVI_621) within the annual calving ground and toz the proportion of calves born on the Arctic National Wildlife Refugez coastal plain physiographic zone where predator density was lowerz than in the foothill-mountain physiographic zone (r 2 = 0.696, P <z 0.001). Calf survival was not estimated in 1986 because inclementz weather prevented a complete sample in late June.z In drder fo assess the potential effects of developmentf of the 1002 Area 5n fhe Porcupine caribou herd duringf calving, we needed a model of caribou behavioralf response fo oil field fhfrastructures. fThe adjacent Centralf Arctic herd (Fig. 3.2), which calved in the vicinity of Prudhoe Bay - Kuparuk complex of petroleumf development areas, provided the only available model of caribou behavioral response to petroleum developmentf during falving.f Parturientfemale cariboubj.e., those about to givef birth or accompanied by very young calves) of the Centralf Arctic herd frepeatedly demostrated fheir sensitivity fof disturbance during the first few weeks df fife of fheirf calves (Smith and Cameron 1983, Whitten and Cameronf 1983, Dau and Cameron 1986; Cameron et al. 1992;f Nellemann and Cameron 1996, 1998).f Parturient females avoided, or fwere less likely tof cross, fnfrastructuresbroads and pipelines) during fhef calving season (Cameron and Whitten 1979, Dau andf Cameron 1986, Murphy and Curatolo 1987, Lawheadf 1988, Cameron ft al. 1992).fIn addition, densities of