This data release includes three datasets used to develop forecasts of autumn body condition for adult female Pacific walruses in the Chukchi Sea during mid and late century time periods. (1) The activity dataset contains daily telemetry records for 218 adult female walruses tracked for periods of 7 to 104 days during 2008-2014, in the Chukchi Sea. Records include the number of hours the walrus was in the water, number of hours the walrus was foraging, study area region where the walrus was located, depths of the foraging locations, and the proportion of the region covered by sea ice. (2) The movement dataset contains telemetry records for 94 of these walruses, giving the dates they moved from one region to another, and the date of the beginning of minimum ice period for that year. (3) The projected-ice dataset contains daily projections of ice conditions in the study area regions derived from 7 general circulation models of future ice availability for mid-century (2045-2054) and late-century (2090-2099) time periods. The movement and activity datasets were developed to model walrus activity and movement as functions of sea ice conditions. The projected-ice dataset was developed to provide input for those models to forecast future walrus activity and movement. Forecasting autumn body condition requires linkage to bioenergetics models.

These data were used to evaluate effects of vessel exposure on Pacific walrus (Odobenus rosmarus divergens) behaviors. We obtained >120,000 hours of location and behavior (foraging, in-water not foraging, and hauled out) data from 218 satellite-tagged walruses and linked them to vessel locations from the marine Automated Information System. This yielded 206 vessel-exposed walrus telemetry hours for comparison to unexposed hours which we used to assess if vessel exposure altered walrus behavior.

These data were used to evaluate effects of vessel exposure on Pacific walrus (Odobenus rosmarus divergens) behaviors. We obtained >120,000 hours of location and behavior (foraging, in-water not foraging, hauled out) data from 218 satellite-tagged walruses and linked them to vessel locations from the marine Automated Information System. This yielded 206 vessel-exposed walrus telemetry hours for comparison to unexposed hours which we used to assess if vessel exposure altered walrus behavior.

Sea ice loss represents a stressor to the Pacific walrus (Odobenus rosmarus divergens), which feeds on benthic macroinvertebrates in the Bering and Chukchi seas. However, no studies have examined the effects of sea ice on foraging walrus space use patterns. Thus, we examined walrus foraging resource selection as a function of proximity to resting substrates and prey biomass with a matched use-availability design. We quantified biomass of 17 benthic taxa, which included amphipods, bivalves, polychaete, sand dollars, tunicates, and sipunculids. We included covariates for distance to sea ice and distance to land, and systematically developed a series of candidate models to examine interactions among benthic prey biomass and resting substrates. We ranked candidate models with Bayesian Information Criterion and made inferences on walrus resource selection based on the top-ranked model. Biomass of the bivalve family Tellinidae, distance to ice, distance to land, and the interaction of distances to ice and land were in the top-ranked model. Standardized model coefficients indicated that distance to ice explained the most variation in walrus foraging resource selection patterns followed by Tellinidae biomass. Distance to land and the interaction of distances to ice and land accounted for similar levels of variation in foraging walrus resource selection. These data represent the used and available resource units with the covariates of distance to land and distance to ice.

An animal's energetic costs are dependent on the amount of time it allocates to various behavioral activities. For Arctic pinnipeds, the time allocated to active and resting behaviors could change with future reductions in sea ice cover and longer periods of open water. The Pacific walrus (Odobenus rosmarus divergens) is a large Arctic pinniped that rests on sea ice or land between foraging trips to feed on the seafloor. This dataset contains behavioral data collected from 216 radio-tagged adult female walruses instrumented in the Chukchi Sea (2008–2014) that formed the basis of a Bayesian generalized linear mixed effects model that investigated probability that a walrus was in water foraging, in water not foraging, or hauled out, as a function of environmental covariates. The data are structured such that each line represents an hour of behavior collected during every third hour of the UTC day (hours 0 through 21), which was associated with temperature and wind metrics derived from the North American Regional Reanalysis dataset, as well as an index of the daily availability of sea ice based on the daily National Ice Center's Marginal Ice Zone product and an index of the availability of land. Each record presents the walrus behavior recorded during the record's hour, the previous hour, and the subsequent hour.

The dataset consists of geospatial files depicting the estimated June-to-November distribution of walrus foraging and occupancy during a four year period of sparse sea ice cover above the Chukchi Sea continental shelf (2008-2011). The walrus distribution and utilization estimates are based on location data from satellite-linked radio-tracked walruses in the Chukchi Sea (2008-2011). Compared to previous years this period was marked by earlier and more extensive sea ice retreat (June - September) and delayed sea ice freeze-up (October - November). This allowed walruses to arrive earlier, occupy slightly more northern areas, and stay later in the Chukchi Sea than previously. Data are combined for all four years and structured by month (June [06] - November [11]). Within each month, we estimated walrus utilization distributions by using an unweighted movement based kernel density estimator. The kernel density was then weighted by the amount of time walruses spent foraging during each recorded track segment. We present contours of the kernel density ranging from the 10th to 95th percentile.

The dataset consists of geospatial files depicting the estimated June-to-November distribution of walrus foraging and occupancy during a four year period of sparse sea ice cover above the Chukchi Sea continental shelf (2008-2011). The walrus distribution and utilization estimates are based on location data from satellite-linked radio-tracked walruses in the Chukchi Sea (2008-2011). Compared to previous years this period was marked by earlier and more extensive sea ice retreat (June - September) and delayed sea ice freeze-up (October - November). This allowed walruses to arrive earlier, occupy slightly more northern areas, and stay later in the Chukchi Sea than previously. Data are combined for all four years and structured by month (June [06] - November [11]). Within each month, we estimated walrus utilization distributions by using an unweighted movement based kernel density estimator. The kernel density was then weighted by the amount of time walruses spent foraging during each recorded track segment. We present contours of the kernel density ranging from the 10th to 95th percentile.

This dataset is comprised of three tables showing the reaction of walruses hauled out on sea ice of the Chukchi Sea to overflights by rotary wing Unoccupied Aircraft Systems (UAS). Data include behavior of walruses before, during, and after overflights, and environmental conditions during flights.

This dataset is comprised of three tables showing the reaction of walruses hauled out on sea ice of the Chukchi Sea to overflights by rotary wing Unoccupied Aircraft Systems (UAS). Data include behavior of walruses before, during, and after overflights, and environmental conditions during flights.

These data are in three folders of Keyhole Markup Language (KML) geospatial polygons representing the outlines of walrus herds apparent in satellite imagery. Each KML file contains one or more geospatial polygons of walrus herd outlines created by one observer who visually interpreted the images. The attribute values from all KML files are collected in a CSV table included with this data package.