Je veux savoir à propos de…

Federal Register / Vol. 86, No. 103 / Tuesday, June 1, 2021 / Proposed Rules
jbell on DSKJLSW7X2PROD with PROPOSALS2

coastal survey flights was as high as 17
percent.
Detailed data on the behavioral responses of polar bears to the aircraft and the distance from the aircraft each polar bear was observed were available for only the flights conducted between 2000 to 2004 n = 581 bears. The AeroCommander 690 was used during this period. The horizontal detection distance from the flight line was recorded for all groups of bears detected. To determine if there was an effect of distance on the probability of a response indicative of potential Level B harassment, we modeled the binary behavioral response by groups of bears to the aircraft with Bayesian probit regression Hooten and Hefley 2019.
We restricted the data to those groups observed less than10 km from the aircraft, which is the maximum distance at which behavioral responses were likely to be reliably recorded. In nearly all cases when more than one bear was encountered, every member of the group exhibited the same response, so we treated the group as the sampling unit, yielding a sample size of 346 groups. Of those, 63 exhibited behavioral responses. Model parameters were estimated using 10,000 iterations of a Markov chain Monte Carlo algorithm composed of Gibbs updates implemented in R R core team 2021, Hooten and Hefley 2019. Normal 0,1
priors, which are uninformative on the prior predictive scale Hobbs and Hooten 2015, were placed on model parameters. Distance to bear as well as squared distance to account for possible non-linear decay of probability with distance were included as covariates. However, the 95 percent credible intervals for the estimated coefficients overlapped zero suggesting no significant effect of distance on polar bears behavioral responses. While it is likely that bears do respond differently to aircraft at different distances, the data available is heavily biased towards very short distances because the coastal surveys are designed to observe bears immediately along the coast. We were thus unable to detect any effect of distance. Therefore, to estimate a single rate of harassment, we fit an interceptonly model and used the distribution of the marginal posterior predictive
probability to compute a point estimate.
Because the data from the coastal surveys were not systematically collected to study polar bear behavioral responses to aircraft, the data likely bias the probability of behavioral response low. We, therefore, chose the upper 99th percentile of the distribution as our point estimate of the probability of potential harassment. This equated to a harassment rate of 0.23. Because we were not able to detect an effect of distance, we could not correlate behavioral responses with profiles of sound pressure levels for the AeroCommander the aircraft used to collect the survey data. Therefore, we could also not use that relationship to extrapolate behavioral responses to sound profiles for takeoffs and landings nor sound profiles of other aircraft.
Accordingly, we applied the single harassment rate to all portions of all aircraft flight paths.
General Approach To Estimating Harassment for Aircraft Activities Aircraft information was determined using details provided in AOGAs Request, including flight paths, flight take-offs and landings, altitudes, and aircraft type. More information on the altitudes of future flights can be found in the Request. If no location or frequency information was provided, flight paths were approximated based on the information provided. Of the flight paths that were described clearly or were addressed through assumptions, we marked the approximate flight path start and stop points using ArcGIS Pro version 2.4.3, and the paths were drawn. For flights traveling between two airstrips, the paths were reviewed and duplicated as closely as possible to the flight logs obtained from www.FlightAware.com FlightAware, a website that maintains flight logs in the public domain. For flight paths where airstrip information was not available, a direct route was assumed. Activities such as pipeline inspections followed a route along the pipeline with the assumption the flight returned along the same route unless a more direct path was available.
Flight paths were broken up into segments for landing, take-off, and traveling to account for the length of time the aircraft may be impacting an
29403

area based on flight speed. The distance considered the landing area is based on approximately 4.83 km 3 mi per 305 m 1,000 ft of altitude descent speed. For all flight paths at or exceeding an altitude of 152.4 m 500
ft, the take-off area was marked as 2.41 km 1.5 mi derived from flight logs found through FlightAware, which suggested that ascent to maximum flight altitude took approximately half the time of the average descent. The remainder of the flight path that stretches between two air strips was considered the traveling area. We then applied the exposure area of 1,610
m 1 mi along the flight paths. The data used to estimate the probability of Level B harassments due to aircraft see section Estimating Harassment Rates of Aircraft Activities suggested 99% of groups of bears were observed within 1.6 km of the aircraft.
We then differentiated the coastal and inland zones. The coastal zone was the area offshore and within 2 km 1.2 mi of the coastline see section Spatially Partitioning the North Slope into coastal and inland zones, and the inland zone was anything greater than 2
km 1.2 mi from the coastline. We calculated the areas in square kilometers for the exposure area within the coastal zone and the inland zone for all takeoffs, landings, and traveling areas. For flights that involve an inland and a coastal airstrip, we considered landings to occur at airstrips within the coastal zone. Seasonal encounter rates developed for both the coastal and inland zones see section Search Effort Buffer were applied to the appropriate segments of each flight path.
Surface encounter rates were calculated based on the number of bears per season see section Search Effort Buffer. To apply these rates to aircraft activities, we needed to calculate a proportion of the season in which aircraft were flown. However, the assumption involved in using a seasonal proportion is that the area is impacted for an entire day i.e., for 24 hours.
Therefore, to prevent estimating impacts along the flight path over periods of time where aircraft are not present, we calculated a proportion of the day the area will be impacted by aircraft activities for each season Table 5.

TABLE 5VARIABLE DEFINITIONS AND CONSTANT VALUES USED IN POLAR BEAR HARASSMENT ESTIMATES FOR WINTER
AND SUMMER AIRCRAFT ACTIVITIES ON THE COAST OF THE NORTH SLOPE OF ALASKA
Variable ds Sp f
VerDate Sep<11>2014

Definition
Value
days in each season
proportion of the season an area of interest is impacted
flight frequency

19:20 May 28, 2021

Jkt 253001

PO 00000

Frm 00041

Fmt 4701

Sfmt 4702

open-water season = 116, ice season = 249
varies by flight.
varies by flight.

E:FRFM01JNP2.SGM

01JNP2