Quiero saber acerca de...

khammond on DSKJM1Z7X2PROD with RULES2

43058

Federal Register / Vol. 86, No. 148 / Thursday, August 5, 2021 / Rules and Regulations
practicable means for required infrared surveys.
Response: Ground-based infrared surveys are not directly comparable to AIR surveys and should not be considered to have equivalent detection rates. We are not aware of any studies that have directly estimated infrared detection from ground-based surveys.
But studies have documented that ground-based infrared is likely to lower detection probability given the greater impacts of blowing snow on detection than when doing aerial surveys Robinson et al. 2014. Pedersen et al.
2020 found that infrared from a vertical position i.e., aerial was four times more likely to detect a den than infrared from a horizontal position i.e., from the ground. Given that our analysis was based on AOGAs proposal to conduct AIR surveys, we did not estimate what the expected level of take would be if ground-based infrared was used instead on a case-by-case basis.
Based on what has been published on the topic, it would not be appropriate to treat ground-based infrared detection as equivalent to aerial-based efforts.
Comment 100: One commenter suggested that the Service should revise their language for the time period in which AIR surveys are to be conducted in order to allow for flexibility due to poor weather and operational complications.
Response: The results of Wilson and Durner 2020 show that specificity in dates when activities occur can significantly affect the level of disturbance expected from industrial activities. The Service worked with AOGA to find date ranges for AIR that met their constraints but that also provided sufficient protection for denning polar bears in light of their proposed activities. AOGA stated they were amenable to these dates.
Unfortunately, the Services analysis is contingent on AIR surveys being conducted within the date ranges in the draft ITR and any deviation from those dates could lead to increased levels of take and harassment of denning bears.
Thus, the Service will not be able to accommodate this request.
Comment 101: One commenter suggested that the Service should clarify whether a third AIR survey is required for seismic survey activities.
Response: The Service is requiring three AIR surveys to occur prior to all seismic activities. The Service has worked together with the applicant to develop mitigation measures that ensure the least practicable adverse impact on polar bears. The applicant agreed that three surveys are practicable and will be
VerDate Sep<11>2014

17:26 Aug 04, 2021

Jkt 253001

conducted prior to all seismic activities as a condition of the LOA.
Comment 102: One commenter suggested that the Service should address whether the reported efficacy of AIR is sufficient to detect polar bear dens prior to the commencement of Industry activities.
Response: We incorporate these different studies cited in the comment that rely on aerial surveys to establish our AIR efficacy used in the model. We dont assume complete detection of dens, but rather a value of 41% with associated uncertainty. So, we account for the inability of an AIR survey to detect all dens in our modeling framework, and the value we use is actually lower than that published in Smith et al. 2020. Robinson et al.
2014 is inappropriate to include because it was based on hand-held ground-based infrared, which is not as effective as aerial surveys, and they do not provide an estimate of detection probability. But we do include the results from Smith et al. 2020 and Amstrup et al. 2004 in our analysis, as well as a new study on artificial dens Woodruff and Wilson 2021.
Comment 103: One commenter suggested that the Service should consider how the variation in weather conditions will affect the efficacy of AIR
to detect polar bear dens.
Response: The AIR efficacy values we use are from a suite of weather conditions and not just optimal conditions, so they cover the range of possible conditions that surveys are flown. For example, Amstrup et al. 2004
found that AIR efficacy was >80% for optimal weather conditions, but we dont use that value. We use the average AIR efficacy, which is closer to 55% for Amstrup. Because a range of weather conditions is used, our estimates are able to provide inference across those conditions. Additionally, while we agree that weather conditions in northern Alaska are likely to change with climate change, surveys are still required to be flown under conditions that have been found to be suitable.
Comment 104: One commenter suggested that the Service should address that polar bear dens can remain undetected despite multiple AIR
surveys in the area and whether the requirement for multiple AIR surveys will effectively increase the den detection rate.
Response: We agree that more AIR
surveys do not make them more effective. Dens in the model can continue to go undetected even after multiple surveys. But, the laws of probability indicate that if you do the surveys multiple times over a den that
PO 00000

Frm 00078

Fmt 4701

Sfmt 4700

is available to be detected, the probability that it will be detected at least once increases. Similar to Amstrup et al. 2004, when you apply two AIR surveys to our simulated dens, the overall probability of detection is only 65%. So, it is incorrect that more surveys do not equal more dens detected.
Comment 105: One commenter suggested that the Service should address how the efficacy of AIR for detecting dens with various depths of snow cover was accounted for in their den detection model.
Response: We dont assume that dens with snow depth >100 cm cant be detected for the current analysis. AIR
efficacy rates are for all dens i.e., independent of snow depth, so by default includes those dens that are unable to be detected for whatever reason. Smith et al. 2020 did not account for snow depth in their detection probability, and Woodruff and Wilson 2021 did not find a relationship between detection and snow depth. That is why we dont take into account snow depth for the approach we took in this model.
Comment 106: One commenter suggested that the Service should clarify their explanation for the sources used to inform their estimation of den detection probability and how uncertainty was accounted for during their estimation of den detection probability.
Response: Our approach is not arbitrary. We are aware of only three studies that utilize AIR detection estimates. Although Scheidler and Perham published a report on aerial survey detections, they had significant issues published in their report that precluded our use of their results. Other studies use drones Pedersen et al. 2020
or handheld infrared Robinson et al.
2014, which are likely not comparable and dont actually provide detection probabilities. With respect to the Woodruff and Wilson 2021 study, the Service published the white paper to give readers details on how the probabilities were derived, but the greater context of the study was not provided because it is currently under peer-review. Many limitations to the study make the use of the lower estimate questionable e.g., onboard navigation equipment was not allowed for observers compared to real surveys.
Thus, we used the detection estimate from that study as the most reliable i.e., dens that were determined to have been covered by the AIR camera. So, the decision was not arbitrary, but based on our in-depth knowledge of the study and its limitations. Lastly, the Service doesnt ignore the uncertainty in den
E:FRFM05AUR2.SGM

05AUR2