Modeling ringed seal Pusa hispida habitat and lair emergence timing in the eastern Bering and Chukchi Seas

Ringed seals Pusa hispida are reliant on snow and sea ice for denning, and a better understanding of ringed seal habitat selection and timing of emergence from snow dens (also called lairs) is needed to quantify and predict effects of climate change in the Arctic. We used generalized additive models to assess relationships between ringed seal counts, from spring aerial surveys in the Bering Sea (2012 and 2013) and Chukchi Sea (2016), and spatiotemporal covariates including survey date, remotely sensed snow and sea-ice values, and short-term weather data. We produced separate models for total ringed seal counts and for pup counts within each region. Our models showed that in both areas, total ringed seal counts increased over the course of the spring, especially after 15 May, indicating emergence from lairs and/or the onset of basking behavior. For the more northerly Chukchi Sea, we found a substantial unimodal effect of snow melt progression and a positive effect of snow depth on total ringed seal counts. In contrast, Bering Sea total ringed seal counts and pup counts in both regions were affected much more strongly by date than by habitat variables. Overall, our findings demonstrate that snow depth and melt play an important role in the timing of ringed seal den emergence, particularly in the Chukchi Sea, and suggest that ringed seal denning may be affected by continued shifts in melt and snow depth associated with climate change.

Modeling haul-out behavior of walruses in Bering Sea ice

Understanding haul-out behavior of ice-associated pinnipeds is essential for designing and interpreting population surveys and for assessing effects of potential changes in their ice environments. We used satellite-linked transmitters to obtain sequential information about location and haul-out state for Pacific walruses, Odobenus rosmarus divergens (Illiger, 1815), in the Bering Sea during April of 2004, 2005, and 2006. We used these data in a generalized mixed model of haul-out bout durations and a hierarchical Bayesian model of haul-out probabilities to assess factors related to walrus haul-out behavior, and provide the first predictive model of walrus haul-out behavior in sea ice habitat. Average haul-out bout duration was 9 h, but durations of haul-out bouts tended to increase with durations of preceding in-water bouts. On average, tagged walruses spent only about 17% of their time hauled out on sea ice. Probability of being hauled out decreased with wind speed, increased with temperature, and followed a diurnal cycle with the highest values in the evening. Our haul-out probability model can be used to estimate the proportion of the population that is unavailable for detection in spring surveys of Pacific walruses on sea ice.

Drilling and operational sounds from an oil production island in the ice-covered Beaufort sea

Recordings of sounds underwater and in air, and of iceborne vibrations, were obtained at Northstar Island, an artificial gravel island in the Beaufort Sea near Prudhoe Bay (Alaska). The aim was to document the levels, characteristics, and range dependence of sounds and vibrations produced by drilling and oil production during the winter, when the island was surrounded by shore-fast ice. Drilling produced the highest underwater broadband (10-10,000 Hz) levels (maximum= 124 dB re: 1 microPa at 1 km), and mainly affected 700-1400 Hz frequencies. In contrast, drilling did not increase broadband levels in air or ice relative to levels during other island activities. Production did not increase broadband levels for any of the sensors. In all media, broadband levels decreased by approximately 20 dB/tenfold change in distance. Background levels underwater were reached by 9.4 km during drilling and 3-4 km without. In the air and ice, background levels were reached 5-10 km and 2-10 km from Northstar, respectively, depending on the wind but irrespective of drilling. A comparison of the recorded sounds with harbor and ringed seal audiograms showed that Northstar sounds were probably audible to seals, at least intermittently, out to approximately 1.5 km in water and approximately 5 km in air.

Tolerance by ringed seals (Phoca hispida) to impact pipe-driving and construction sounds at an oil production island

During June and July 2000, impact pipe-driving sounds at Northstar Island (Prudhoe Bay, Alaska) were recorded underwater and in air at distances 63-3000 m from the source. Simultaneously, reactions of nearby ringed seals (in water or on ice) were documented. Pipe-driving pulses were analyzed to determine unweighted peak and rms sound-pressure levels (SPLs) and sound-exposure levels (SELs). Underwater, mean levels for these parameters reached 157 and 151 dB re: 1 microPa and 145 dB re: 1 microPa2 x s, respectively, at 63 m. The corresponding values in air were 112 and 96 dB re: 20 microPa and 90 dB re: (20 microPa)2 x s, respectively. Underwater SPLs were <180 dB re: 1 microPa at all distances. During 55 h of observation, 23 observed seals exhibited little or no reaction to any industrial noise except approaching Bell 212 helicopters. Ringed seals swam in open water near the island throughout construction activities and as close as 46 m from the pipe-driving operation. Based on current audiometric data for seals, these sounds are expected to be audible to less than 3 km underwater and at least 0.5 km in air. Most likely the seals around Northstar Island were habituated to industrial sounds.