Voglio sapere a riguardo di…

khammond on DSKJM1Z7X2PROD with RULES2

43060

Federal Register / Vol. 86, No. 148 / Thursday, August 5, 2021 / Rules and Regulations
Additionally, in the Woodruff and Wilson study, an attempt was made to reduce the influence of subsequent snowfall on their ability to estimate the relationship. Thus, they restricted their assessment of snow depth to data from the first two surveys, to minimize the effects of additional snowfall accumulation on the relationship with detection. We disagree that the study by Woodruff and Wilson suggests AIR
surveys are unlikely to detect dens.
Instead, the study shows that dens can be detected, but that the efficacy may not be as high as previously thought.
Hence, they actually state unlikely to detect all dens. However, we take this into account in our analysis and consider the range of studies that have addressed AIR efficacy to derive AIR
efficacy used in this analysis. Also see Optimal Weather Conditions for AIR.
Comment 115: One commenter suggested that the Service should consider requesting additional AIR
surveys to increase their polar bear den detection rate.
Response: We rely on the best available science regarding the efficacy of aerial IR surveys to detect established polar bear dens. The Service considered multiple options for the number of AIR
flights that would be required and found the number published in the ITR
adequate for reducing take sufficiently while still feasible for Industry to conduct during the short period of the winter when AIR flights can be reliably done.
Comment 116: One commenter suggested that the Service overestimated their polar bear den detection rate for AIR surveys based on comparisons to other studies.
Response: The 74% detection estimate was only for dens with snow depth <100 cm. The actual probability when ignoring snow depth is closer to 55%. Wilson and Durner are clear how they derived these estimates, and they rely on standard probability methods.
The commenter states that the appropriate metric to use from Amstrup et al. 2004 is the overall number of dens detected at least once, divided by the total number of dens surveyed. This is inappropriate, however, because it doesnt take into account how many times each den was surveyed. With increasing search effort, the probability of detecting the den increases. Wilson and Durner 2020 obtained an estimate for the probability of detecting a den on a single survey. We disagree that our mathematic approach defies logic, or that the approach is sophisticated its a simple binomial model because some dens were detected on multiple occasions. Thus, it is inappropriate to
VerDate Sep<11>2014

17:26 Aug 04, 2021

Jkt 253001

ignore those detections by simply looking at the overall number of dens detected at least once. Wilson and Durner 2020 provide a thorough explanation of their approach, and it went through numerous rounds of peerreview during which the method was deemed appropriate and published in a well-respected peer-reviewed journal.
Comment 117: Two commenters suggested that the Service should clarify their explanation for averaging the polar bear den detection rates across multiple studies considering that each study was conducted in a different situation that may not be applicable to the AIR
surveys required for these regulations.
Response: We disagree. Our approach actually tries to accommodate the different limitations of each of the studies. Amstrup was based on real dens; Woodruff was based on artificial dens. Amstrup used all of the tools an aircraft had to offer, but Woodruff didnt to try and control observer learning where dens were. Smith didnt include any estimate of uncertainty in their estimate of detection, nor did they provide much information on methods or underlying data and search effort. So, each study has their own set of limitations, but to our knowledge, these are the only data from AIR surveys. So, it represents the best available information. And the overall probability used wasnt significantly different from the results of Smith et al. 2020.
Comment 118: One commenter suggested that the Service should clarify their explanation for their polar bear den detection rate and recommended that this detection rate is unlikely to exceed 50%.
Response: Unfortunately, 50% is not an appropriate metric from the Woodruff and Wilson study because it doesnt account for search effort nor multiple detections of a den across surveys. Both of those factors need to be considered when estimating a detection probability. Additionally, the public was given only a brief overview of the study and methods because the actual study is still under journal peer review.
Thus, the commenter doesnt have all of the caveats associated with that study and why it is likely inappropriate to consider only the results from the Woodruff and Wilson study and ignore the previously published works that have different study designs and pros/
cons. The detection probability we derived and used from Amstrup et al. is also very similar to Smith et al. We agree with the commenter that it is unlikely that real-world AIR will have a detection >50%. As we showed in our analysis, our mean detection was 41%

PO 00000

Frm 00080

Fmt 4701

Sfmt 4700

and could go as low as 1.5% during any given iteration of the model.
Comment 119: One commenter suggested that the Service should specify flight paths for AIR surveys to ensure complete coverage of all polar bear denning habitat and require that AIR surveys conduct multiple passes across denning habitat as well as use helicopters for AIR surveys to increase den detection rates.
Response: Because the estimates used for AIR efficacy are based on the range of suitable weather conditions under which AIR surveys are acceptable, and the analytical approach requires that all den habitat as identified in the studies cited is adequately surveyed, the ITR
already implicitly requires these to occur.
Comment 120: One commenter suggested that the Service should address how the efficacy of AIR for detecting polar bear dens with more than 90 cm of snow cover was accounted for in their den detection model.
Response: While one study Robinson et al. 2014 showed lower detectability for dens in snow deeper than 90 cm, it was based on handheld infrared, not aerial. And in the Woodruff and Wilson study, a den with snow 145 cm deep was detected, so a simple cutoff based on ground-based infrared is likely not appropriate.
Comment 121: One commenter suggested that the Service should consider the practicality of requesting Industry entities to complete three AIR
surveys prior to commencing activities in polar bear denning habitat.
Response: The ITR does not indicate that industry will conduct three AIR
surveys of all habitat. Three surveys are required only for areas receiving seismic surveys, and in years when seismic occurs, along the pipeline corridor between Deadhorse and Pt. Thomson.
Regardless, our analysis requires that all den habitat within 1 mile of industrial activity/infrastructure will receive at least two AIR surveys under conditions suitable for detecting dens. Only if industry flies all of the AIR surveys required per the analysis will they have coverage under the ITR.
The Service notes that the extent of AIR surveys required by this ITR
significantly exceeds what has been required under prior iterations of the ITR and is sufficient to ensure that all applicable MMPA standards are met, including the requirement to prescribe means to effect the least practicable adverse impact on the species or stock and its habitat.
Comment 122: One commenter suggested that the Service should
E:FRFM05AUR2.SGM

05AUR2