We provide here the data used in analysis of 3 test cases, presented in the manuscript "Ecological insights from three decades of animal movement tracking across a changing Arctic". We utilized the new Arctic Animal Movement Archive (AAMA), a growing collection of 201 standardized terrestrial and marine animal tracking studies from 1991–present. The AAMA supports public data discovery, preserves fundamental baseline data for the future, and facilitates efficient, collaborative data analysis. With three AAMA-based case studies, we document climatic influences on the migration phenology of eagles, geographic differences in adaptive response of caribou reproductive phenology to climate change, and species-specific changes in terrestrial mammal movement rates in response to increasing temperature.  

Arctic Animal Movement Archive (AAMA)

The AAMA enables the discovery and use of Arctic and Subarctic animal movement data. The AAMA is hosted on Movebank (movebank.org), a global research platform for animal tracking and bio-logging data. Current information about participating studies, how to download or request data, and how to join the archive and provide new datasets is available at the AAMA landing page (https://www.movebank.org/cms/movebank-content/arctic-animal-movement-archive). 

The data presented here is processed from raw movement data in AAMA. This data publication provide the data used in 3 case studies that analyze ecological changes in animal movement, phenology and demography in response to environmental conditions in the arctic.

Case Study 1. Long-term data reveal influence of decadal climate patterns on summering behavior in a migratory raptor

We used 569,720 location estimates collected from 146 golden eagles (Aquila chrysaetos) in 12 AAMA studies across western North America (Table 1). Records were imported into R via the move package. We assigned all individuals to one of three age categories: juvenile (survived up to the onset of its 2^nd southbound migration), sub-adult (surviving up to the onset of its 4^th southbound migration), or adult (surviving past the onset of its 4^th southbound migration). We incorporated into analyses instances where individuals were observed during an age-class transition. Breeding status was not considered in analyses because it was not determined for all individuals. We identified the onset of golden eagle summering behavior via a mechanistic range shift analysis in R with the marcher package. The model was fit with the maximum likelihood method, and we used the Ornstein-Uhlenbeck model option to estimate ranging behavior. We modified the MRSA slightly by forcing the model to fail if a data gap of >30 days occurred. We inspected start dates and locations to remove spurious model results due to tag deployments that ended during spring migration or began during summer. This process yielded 179 summering onset estimates with modelled start and end dates made by 103 eagles.

Table 1.

Studies used in Case Study 1, ‘long-term data reveal importance of decadal climate patterns on summering behavior in a migratory raptor’.

 

Case Study 2. Large-scale geographic differences in parturition timing of caribou

We used caribou movement data in 13 AAMA studies (Table 2). Among the caribou in our study, woodland caribou (Rangifer tarandus caribou) are considered a separate subspecies from barren-ground caribou (R. t. groenlandicus), which are known for their long-distance calving migrations. Woodland caribou are further classified as belonging to different ecotypes, including boreal and mountain, which differ by habitat and behavioral phenotype. Following exploratory analysis, we further partitioned woodland caribou by geographic area into northern and southern mountain and boreal caribou. Barren-ground caribou data were partitioned into their largely coherent reproductive subpopulations (herds).

Parturition dates were estimated by adapting methods described in DeMars et al. (2013) for each of the two subspecies. For R. t. caribou, we analyzed movements between April 28 and June 30 and found times when movement activity decreased suddenly and persistently below 5 threshold speeds (10, 15, 20, 25, and 30 m/h), and conservatively retained those parturitions where three of five thresholds agreed within a day of each other. For barren-ground caribou, which calve significantly later and shortly after completing a long-distance migration, we adapted the individual-based method (DeMars et al. 2013; Gurarie et al. 2019), which fits a non-calving model, in which movement rates remain constant throughout the study period, and a calving model, where there is a sudden drop in movement rate followed by a progressive increase.

 

Table 2.

Studies used in Case Study 2, ‘large-scale geographic differences in parturition timing of caribou’.

 

Case Study 3: Temperature and precipitation response in movement rates of terrestrial mammals

We compiled movement paths for 1,720 individuals representing five mammalian species from 22 AAMA studies: black bear, Ursus americanus; grizzly bear, U. arctos; caribou, R. tarandus ssp.; moose, Alces alces; and wolf, Canis lupus (Table 3). We calculated step lengths (m) and location sampling interval (min) along each movement time series, excluding locations with sampling intervals longer than 24 hours. We annotated movement data with daily maximum temperatures (°C) and summertime precipitation or winter snow water equivalent (SWE) estimates using the Daymet data product (https://daymet.ornl.gov/). To match the temporal scale of Daymet, we aggregated movements by estimating daily mean step lengths and sampling intervals for each day within an individual’s time series. In order to reduce the influence of seasonally dependent behaviors (e.g., migration or parturition), we partitioned data into January-only (peak winter) and July-only (peak summer) datasets. As in case study 2, we treated the two subspecies of caribou (R. t. groenlandicus and R. t. caribou) as distinct groups within our analysis due to anticipated differences in daily movement rates. We also excluded bears from the winter analyses due to lack of data; however, the very nature of their hibernation represents a reduced overwinter movement rate.

 

Table 3.

Studies used in Case Study 3, ‘temperature and precipitation response in movement rates of terrestrial mammals’.

 

Data files for models associated with analyses reported in Davidson et al. (2020) "Ecological insights from three decades of animal movement tracking across a changing Arctic". As relevant, files follow the vocabulary used at movebank.org:
Max Planck Institute of Animal Behavior (2019) Movebank Attribute Dictionary, Version 2. Natural Environment Research Council. http://vocab.nerc.ac.uk/collection/MVB

#############
eagles_summering_onsets.csv: This file contains data from the case study on long-term influence of decadal climate patterns on summering behavior in a migratory raptor. Records represent modeled summering behavior onsets golden eagles. Onset locations and dates were determined via a mechanistic range shift analysis, using 569,720 location estimates of 146 golden eagles. Attributes include

animal-taxon: The scientific name of the species on which the tag was deployed, as defined by the Integrated Taxonomic Information System (ITIS, www.itis.gov).
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000024

animal-id: An individual identifier for the animal, provided by the data owner. Units: none
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000016

timestamp: The date and time of the summering onset date. Format: yyyy-MM-dd HH:mm:ss.SSS, in UTC time.

location-lat: The geographic latitude of the summering onset location. Units: decimal degrees, WGS84 reference system
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000145

sensor-type: The type of sensor with which location data were collected.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000170

animal-sex: The sex of the animal. Allowed values are m = male; f = female; u = unknown.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000023 

animal-life-stage: Age class category for the animal-year assessed. For each animal, age at the first year corresponds with the animal-life-stage at time of deployment in the Movebank study, and is adjusted if needed for subsequent years to represent age transitions as described in Supplemental Materials.

previous-year: The year of the previous winter season for which Pacific Decadal Oscillation data were assessed. Equal to summering behavior year - 1.

winter-PDO: averaged mean monthly Pacific Decadal Oscillation data for the previous winter (i.e., November – March). Data from http://research.jisao.washington.edu/pdo

study-name: The name of the study participating in the AAMA on Movebank (movebank.org) in which the location data from which summering onsets were estimated are stored.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000173

#############
caribou_parturitions.csv: This file contains data from the case study on large-scale geographic differences in parturition timing of barren-ground (Rangifer tarandus groenlandicus) and woodland (R. tarandus caribou) caribou. Records represent daily mean location for the day of estimated parturition.

animal-taxon: The scientific name of the species on which the tag was deployed, as defined by the Integrated Taxonomic Information System (ITIS, www.itis.gov).
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000024

animal-taxon-detail: A more specific name and/or reference for the taxon name provided by 'animal taxon'. This can be used, for example, to specify a subspecies or a taxon not supported by the ITIS.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000025

animal-id: An individual identifier for the animal, provided by the data owner. Units: none
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000016

animal-nickname: An alternate identifier for the animal.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000020

population: The population as referred to in the paper text.

barren-ground herd: The herd name for barren-ground caribou as referred to in the paper text. Note that the herd assignment for an individual does not always correspond to the Movebank study with the same name.

timestamp: The date and time of the parturition date. Format: yyyy-MM-dd HH:mm:ss.SSS, in UTC time.

doy: The Julian day of year.

year: The year.

height-above-msl: The estimated height of the tag above mean sea level, typically estimated by the tag. Units: meters
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000131
For consistency across studies, elevation data are taken from Canadian Digital Elevation Data at https://open.canada.ca/data/en/dataset/7f245e4d-76c2-4caa-951a-45d1d2051333 (Natural Resources Canada).

x.km: The east-west component of the parturition location, represented as the distance from the median location within each “population” group.

y.km: The north-south component of the parturition location, represented as the distance from the median location within each “population” group.

study-name: The name of the study participating in the AAMA on Movebank (movebank.org) in which the location data from which parturitions were estimated are stored.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000173

#############
mammal_movement_rates.csv: This file contains data from the case study on temperature and precipitation response in movement rates of terrestrial mammals. Records represent daily movement rates for individuals calculated from original location estimates.

animal-taxon: The scientific name of the species on which the tag was deployed, as defined by the Integrated Taxonomic Information System (ITIS, www.itis.gov).
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000024

animal-taxon-detail: A more specific name and/or reference for the taxon name provided by 'animal taxon'. This can be used, for example, to specify a subspecies or a taxon not supported by the ITIS.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000025

animal-id: An individual identifier for the animal, provided by the data owner. Units: none
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000016

year: The year.

season: The season (summer represented by July or winter represented by January).

doy: The day of year.

mnSL: Daily mean step length (meters).

mnDT: Daily mean sampling interval (minutes).

stdTemp: The daily maximum temperature (degrees Celsius), centered on means and scaled by 1 standard deviation. Temperature data from Daymet (http://daymet.ornl.gov).

stdSWE: The snow water equivalent (kg•m^-2), centered on means and scaled by 1 standard deviation. SWE data from Daymet (http://daymet.ornl.gov).

stdPrecip: The daily cumulative precipitation (mm), centered on means and scaled by 1 standard deviation. Precipitation data from Daymet (http://daymet.ornl.gov).

study-name: The name of the study participating in the AAMA on Movebank (movebank.org) in which the location data from which movement metrics were calculated are stored.
http://vocab.nerc.ac.uk/collection/MVB/current/MVB000173