Voglio sapere a riguardo di…

khammond on DSKJM1Z7X2PROD with RULES2

43026

Federal Register / Vol. 86, No. 148 / Thursday, August 5, 2021 / Rules and Regulations
if a behavioral response was observed.
However, multiple exposures do not result in multiple Level B harassment takes unless the exposures occurred in two different denning periods.
If comprehensive dates of specific exposures are not available and daily exposures were possible e.g., the den was located within 1.6 km 1 mi of an ice road, we assumed exposures occurred daily.
In the event of an exposure that resulted in a disturbance to denning bears, take was assigned for each bear i.e., female and each cub associated with that den. Whereas assigned take for cubs could range from Level B
harassment to lethal take, for adult females only Level B harassment was possible.
In the absence of additional information, we assumed dens did not contain cubs prior to December 1 but did contain cubs on or after December 1.
If an exposure occurred and the adult female subsequently abandoned her cubs, we assigned a lethal take for each cub.
If an exposure occurred during the early denning period and bears emerged from the den before cubs reached 60
days of age, we assigned a lethal take for each cub. In the absence of information about cub age, a den emergence that occurred between December 1 and February 13 was considered to be an early emergence and resulted in a lethal take of each cub.
If an exposure occurred during the late denning period i.e., after cubs reached 60 days of age and bears emerged from the den before their intended i.e., undisturbed emergence date, we assigned a serious injury Level A harassment take for each cub. In the absence of information about cub age and intended emergence date which was known only for simulated dens, den emergences that occurred between and including February 14 and March 14 were considered to be early emergences and resulted in a serious injury Level A harassment take of each cub. If a den emergence occurred after March 14 but was clearly linked to an exposure e.g., bear observed emerging from the den when activity initiated near the den, we considered the emergence to be early and resulted in a serious injury Level A harassment take of each cub.
For dens where emergence was not classified as early, if an exposure occurred during the post-emergence period and bears departed the den site prior to their intended i.e., undisturbed departure date, we assigned a non-serious injury Level A

VerDate Sep<11>2014

17:26 Aug 04, 2021

Jkt 253001

harassment take for each cub. In the absence of information about the intended departure date which was known only for simulated dens, den site departures that occurred less than 8
days after the emergence date were considered to be early departures and resulted in a non-serious injury Level A
harassment take of each cub.
Den Simulation We simulated dens across the entire north slope of Alaska, ranging from the areas identified as denning habitat Blank 2013, Durner et al. 2006, 2013
contained within the National Petroleum ReserveAlaska NPRA in the west to the Canadian border in the east. While AOGAs Request does not include activity inside ANWR, we still simulated dens in that area to ensure that any activities directly adjacent to the refuge that might impact denning bears inside the refuge would be captured. To simulate dens on the landscape, we relied on the estimated number of dens in three different regions of northern Alaska provided by Atwood et al. 2020. These included the NPRA, the area between the Colville and Canning Rivers CC, and ANWR.
The mean estimated number of dens in each region during a given winter were as follows: 12 dens 95% CI: 326 in the NPRA, 26 dens 95% CI: 1148 in the CC region, and 14 dens 95% CI: 5
30 in ANWR Atwood et al. 2020. For each iteration of the model described below, we drew a random sample from a gamma distribution for each of the regions based on the above parameter estimates, which allowed uncertainty in the number of dens in each area to be propagated through the modeling process. Specifically, we used the method of moments Hobbs and Hooten 2015 to develop the shape and rate parameters for the gamma distributions as follows: NPRA 122/5.82,12/5.82, CC
262/9.52,26/9.52, and ANWR 142/
6.32,14/6.32.
Because not all areas in northern Alaska are equally used for denning and some areas do not contain the requisite topographic attributes required for sufficient snow accumulation for den excavation, we did not randomly place dens on the landscape. Instead, we followed a similar approach to that used by Wilson and Durner 2020 with some additional modifications to account for differences in denning ecology in the CC
region related to a preference to den on barrier islands and a general but not complete avoidance of actively used industrial infrastructure. Using the USGS polar bear den catalogue Durner et al. 2020, we identified polar bear dens that occurred on land in the CC

PO 00000

Frm 00046

Fmt 4701

Sfmt 4700

region and that were identified either by GPS-collared bears or through systematic surveys for denning bears Durner et al. 2020. This resulted in a sample of 37 dens of which 22 i.e., 60
percent occurred on barrier islands. For each iteration of the model, we then determined how many of the estimated dens in the CC region occurred on barrier islands versus the mainland.
To accomplish this, we first took a random sample from a binomial distribution to determine the expected number of dens from the den catalog Durner et al. 2020 that should occur on barrier islands in the CC region during that given model iteration; nbarrier =
Binomial37,22/37, where 37 represents the total number of dens in the den catalogue Durner et al. 2020 in the CC
region suitable for use as described above and 22/37 represents the observed proportion of dens in the CC
region that occurred on barrier islands.
We then divided nbarrier by the total number of dens in the CC region suitable for use i.e., 37 to determine the proportion of dens in the CC region that should occur on barrier islands i.e., pbarrier. We then multiplied pbarrier with the simulated number of dens in the CC
region rounded to the nearest whole number to determine how many dens were simulated to occur on barriers islands in the region.
In the NPRA, the den catalogue Durner et al. 2020 data indicated that two dens occurred outside of defined denning habitat Durner et al. 2013, so we took a similar approach as with the barrier islands to estimate how many dens occur in areas of the NPRA with the den habitat layer during each iteration of the model; nhabitat
Binomial15, 13/15, where 15
represents the total number of dens in NPRA from the den catalogue Durner et al. 2020 suitable for use as described above, and 13/15 represents the observed proportion of dens in NPRA
that occurred in the region with den habitat coverage Durner et al. 2013. We then divided nhabitat by the total number of dens in NPRA from the den catalogue i.e., 15 to determine proportion of dens in the NPRA region that occurred in the region of the den habitat layer phabitat.
We then multiplied phabitat with the simulated number of dens in NPRA
rounded to the nearest whole number to determine the number of dens in NPRA that occurred in the region with the den habitat layer. Because no infrastructure exists and no activities are proposed to occur in the area of NPRA without the den habitat layer, we only considered the potential impacts of activity to those dens simulated to occur
E:FRFM05AUR2.SGM

05AUR2