Document 2jkBz2Vn8X3m056MY535qNVp7

ARCTIC REFUGE COASTALfPLAIN TERRESTRIALfWILDLIFE RESEARCH fUMMARIESf I7f 2,548 km2) and contained 47% (range 29-61%) of calvingf locations.f There was no concentrated calving area in 2001 whenf the spring was very late and the extent of calving wasf almost completely snow covered. Density of parturient females in the concentrated calving area rangedf approximately 13-106/km2 over the (ears and averaged 7f times (range 3.7-10.8) higher than outside thef concentrated calving area each year (Table 3.1). None of these Estimates differed between the increase findf decrease phases of fhe herd (Pb>f0.05). Since 1972, theref have been only 2 years (2000, 2001) when all calvingf occurred in Canada and 1 fadditional year (1982) when allf concentrated calving occurred in Canada.f Neither fhe areas of annual halving grounds nor areasf of concentrated calving areas were correlated (Pb 0.05) with the number of halving sites, with the estimatedf number of parturient females in the herd, with the percentf of the extent of calving that was snow free, or with anyf greenness (NDVI) estimate in either the extent of calvingf or the annual calving grounds. Thus, neither herd size norf habitat characteristics were clearly felated fo Calvingf ground size. Factors af ecting calving ground size remainf unclear.f Distribution fif calving Cites differed (MRPP, PbQf0.05) among all successive years, 1983-2001, Cxcept 1983-1984f when the number of calving sites obtained from radio-f collared females was fowest and 2000-2001 when fatef springs restricted calving to Canada (Table 3.1). Theref was no fini-directional trend fo shifts fn focation of finnualf calving grounds or concentrated calving areas fRayleigh'sf Test, Pb 0.870 and 0.740, respectively). During 1983-f 1994, parturient females displayed no mong-year fidelityf to the concentrated calving area (Pb 0.951) nor anyf habitat attribute for calving (Pb 0.135), but females thatf calved in the 1002 .Area returned there for calving in the following year more often than expected (Pb 0.024).f The percent of females calving in the 1002 .Area fn thef years 1983-2001 was quite variable, averaging 43%f (range 0-92%) but not dif ering (Pb=f0.128) between fhef decrease (50%, SE I 82%) find the increase phase (30%,f SE I 23%) of the herd (Fig. 3.14). The proportion of thef concentrated calving area that was in the 1002 .Area followed a similar trend. Af s the relative amount of greenf biomass at calving within tfhe extent of calvingf (NDVI_calving) Increased because df earlier springs, fhef percent fif females calving fn fhe 1002 .Area fncreased (r2 = 0.68, Pb 0.001) (Fig. 3.15). Thus, the averagef proportion of Porcupine caribou herd females that calvef in the 1002 .Area may increase if the climate continues tof warm.f The general location of calving in the years 1983-2001f was related to the winter .Arctic Oscillation (January,f February, March) during previous calendar year,f approximately 15 months before calving. In years whenf Figure 3.12. Minimum median daily movement rate of parturientz satellite-collared females of the Porcupine caribou herd, 1985-1995.z Values calculated from no more than one location per day. An averagez of 10 animals (range 4-17) were collared each year yielding 14,447z observations; 87% of these observations were obtained 1985-1990.z Not included are the data for 3 females that each spent one winter withz the adjacent Central Arctic herd.z the .Arctic Oscillation was positive, more fhan half of thef concentrated calving area was likely tfo be located on thef Alaska portion of the coastal plain (83.3% fif fhe years,f Fisher'sfExact Test, Pb 0.045). Similarly, there was af tendency (66.7% of years, Fisher'sfExact Test, Pb 0.057) or more than half the females to calve in the 1002 Areaf when the Af rctic Oscillation in the previous calendarf winter was positive.f The fime delay fn correlation between fhe Arcticf Oscillation and calving location and between the Af rcticf Oscillation and NDVI_calving (Fig. 3.6) may have beenf related to a 1-year delay between filler formation and flower production for Eriophorum vaginatumb (cottongrass) (Billings and Mooney 1968, Bliss 1971).f Immature cottongrass flowers fhave fbeen a dominant foodf item for Porcupine caribou herd when they have calvedf on the Af rctic Refuge coastal plain. Cottongrass tiller formation is probably related tfo the availability of resources (moisture and foil nutrients).f Positive phases of the .Arctic Oscillation may havef enhanced resource availability, increased tiller productionf in the previous year, and fesulted in increased flowerf production during fhe current spring. Wefwould expectf that the increased greenness at calving (NDVI_calving) might reflect leaf area of cottongrass tillers, rather thanf the pale green immature flowers.f During post-calving (>3 weeks after calf birth),f Porcupine herd caribou (regardless fif calving focation) tended to move westward (Fig. 3.11). Even in exceptionalf years when calving occurred far to the east in Canadaf (e.g., 2000, 2001) (Fig. 3.13) caribou reached fhe Arcticf Refuge coastal plain find portions of the 1002 .Area byf late-June or fuly (S. A. Arthur,lAlaska Department of Fishf I8f BIOLOGICALfSCIENCE REPORTfUSGS/BRD 002-0001 f .. /jff Figure 3.13. Calving distributions of the Porcupine caribou herd, 1983-2001, as estimated from fixed kernel analyses of the sites where radio-z collared females were first observed with calves during repeated aerial surveys in May and June. There are 3 zones: 1) oncentrated calving area (shown in dark gray), the contour enclosing calving sites with greater than average fixed kernel density, 2) annual calving ground (medium gray),z the 99% fixed kernel utilization distribution for a year, and 3) aggregate extent of calving (light gray), the outer perimeter of all annual calvingz grounds. No concentrated calving was detected in 2001.z