ENDANGERED SPECIES RESEARCHEndang Species Res

Vol. 32: 19–33, 2017doi: 10.3354/esr00779

Published January 12

INTRODUCTION

Knowledge of an animal’s behavior can inform species conservation and management by revealinghow individuals respond to environmental conditions

(Suther land 1998, Caro 1999, Cooke et al. 2014).Although visual observation is the most directmethod to study animal behavior, it is impractical formany species. Innovations in electronic logging andtracking devices have provided new methods to

© The authors 2017. Open Access under Creative Commons byAttribution Licence. Use, distribution and reproduction are un -restricted. Authors and original publication must be credited.

Publisher: Inter-Research · www.int-res.com

*Corresponding author: apagano@usgs.gov

Using tri-axial accelerometers to identify wildpolar bear behaviors

A. M. Pagano1,2,*, K. D. Rode1, A. Cutting3, M. A. Owen4, S. Jensen5, J. V. Ware6, C. T. Robbins7, G. M. Durner1, T. C. Atwood1, M. E. Obbard8, K. R. Middel8,

G. W. Thiemann9, T. M. Williams2

1US Geological Survey, Alaska Science Center, 4210 University Dr., Anchorage, AK 99508, USA2Department of Ecology & Evolutionary Biology, University of California, Santa Cruz, CA 95060, USA

3Oregon Zoo, Portland, OR 97221, USA4Institute for Conservation Research, San Diego Zoo Global, San Diego, CA 92027, USA

5Alaska Zoo, Anchorage, AK 99507, USA6Department of Integrative Physiology and Neuroscience, Washington State University, Pullman, WA 99164, USA

7School of the Environment and School of Biological Sciences, Washington State University, Pullman, WA 99164, USA8Wildlife Research and Monitoring Section, Ontario Ministry of Natural Resources and Forestry, Trent University,

Peterborough, ON K9L 0G2, Canada9Faculty of Environmental Studies, York University, Toronto, ON M3J1P3, Canada

ABSTRACT: Tri-axial accelerometers have been used to remotely identify the behaviors of a widerange of taxa. Assigning behaviors to accelerometer data often involves the use of captive animalsor surrogate species, as their accelerometer signatures are generally assumed to be similar tothose of their wild counterparts. However, this has rarely been tested. Validated accelerometerdata are needed for polar bears Ursus maritimus to understand how habitat conditions may in -fluence behavior and energy demands. We used accelerometer and water conductivity data toremotely distinguish 10 polar bear behaviors. We calibrated accelerometer and conductivity datacollected from collars with behaviors observed from video-recorded captive polar bears andbrown bears U. arctos, and with video from camera collars deployed on free-ranging polar bearson sea ice and on land. We used random forest models to predict behaviors and found strong ability to discriminate the most common wild polar bear behaviors using a combination ofaccelerometer and conductivity sensor data from captive or wild polar bears. In contrast, modelsusing data from captive brown bears failed to reliably distinguish most active behaviors in wildpolar bears. Our ability to discriminate behavior was greatest when species- and habitat-specificdata from wild individuals were used to train models. Data from captive individuals may be suitable for calibrating accelerometers, but may provide reduced ability to discriminate somebehaviors. The accelerometer calibrations developed here provide a method to quantify polarbear behaviors to evaluate the impacts of declines in Arctic sea ice.

KEY WORDS: Activity · Behavior · Polar bear · Ursus maritimus · Acceleration · Accelerometer ·Brown bear · Ursus arctos

OPENPEN ACCESSCCESS

Endang Species Res 32: 19–33, 2017

study the behavior, movement, physiology, energeticrates, and environmental conditions of wildlife thatmay otherwise be difficult or impossible to monitor(Ropert-Coudert & Wilson 2005, Cooke 2008, Wilsonet al. 2008, Bograd et al. 2010, Costa et al. 2010,Wilmers et al. 2015).

Tri-axial accelerometers, which collect high fre-quency measures of acceleration in the form of gra -vitational and inertial velocity (Brown et al. 2013),have provided a means to remotely identify animalbehaviors (Yoda et al. 1999, Watanabe et al. 2005).Accelerometers have been particularly useful instudying widely dispersed animals or those occurringin remote habitats, such as marine mammals andbirds (Brown et al. 2013). Once calibrated, tri-axialaccelerometer data from wild animals can be used toremotely identify behaviors such as resting, walking,running, and even feeding events (Yoda et al. 2001,Shepard et al. 2008, Wilson et al. 2008, Watanabe &Takahashi 2013, Williams et al. 2014). Calibrationtypically involves time-synchronizing behavioral ob -servations with their associated accelerometer read-ings, which often necessitates the use of captive ani-mals or surrogate species (e.g. Yoda et al. 2001,Shepard et al. 2008, Nathan et al. 2012, Campbell etal. 2013). Alternatively, animal-borne video camerascan be used to directly calibrate accelerometers (e.g.Watanabe & Takahashi 2013, Nakamura et al. 2015,Volpov et al. 2015), but cameras can be expensiveand can only collect data over limited durations.

Polar bears Ursus maritimus typically occupy re -mote environments, and few quantitative data existon their behaviors or activity budgets. Much of whatis known about polar bear behavior on the sea icecomes from coastal indigenous resident knowledge(e.g. Nelson 1966, Kalxdorff 1997, Kochnev et al.2003, Voorhees et al. 2014) and direct observationalresearch limited to 2 locations over limited time pe -riods (Stirling 1974, Stirling & Latour 1978, Hansson& Thomassen 1983, Stirling et al. 2016). Satellitetelemetry has been used to track polar bears in somesubpopulations since the late 1970s (Schweinsburg &Lee 1982, Taylor 1986) and has helped to identifyimportant habitats (Ferguson et al. 2000, Mauritzenet al. 2003, Durner et al. 2009, Wilson et al. 2014).However, detailed behavioral data in associationwith habitat conditions are lacking. Recent declinesin Arctic sea ice have already caused declines inabundance, survival, or body condition of polar bearsin some subpopulations (Stirling et al. 1999, Regehret al. 2007, Rode et al. 2010, 2012, Bromaghin et al.2015, Obbard et al. 2016) and models project increas-ing negative impacts in the 21st century (Amstrup et

al. 2008, Hunter et al. 2010, Molnár et al. 2010,Atwood et al. 2016). In order to better predict theimpacts of projected sea ice loss on polar bears, it willbe important to understand the behavioral andphysio logical mechanisms driving current declines(Vongraven et al. 2012, Atwood et al. 2016). Ac -celerometers could be used in combination withsatellite telemetry to better understand the behav-ioral consequences of sea ice loss. This mechanisticinformation would allow for improved assessment ofthe relationships between habitat loss, individualhealth, and vital rates in polar bear populations.

In this study, we developed a method to quantifywild polar bear behaviors using accelerometers andconductivity sensor data, validated through animal-borne video camera data. Additionally, we evaluatedthe effectiveness of using accelerometer data fromcaptive polar and brown bears U. arctos to predictbehaviors of wild polar bears. Though it is generallyassumed that accelerometer signatures of captives orsurrogates are similar to those of their instrumentedwild counterparts (Williams et al. 2014, McClune etal. 2015, Wang et al. 2015, Hammond et al. 2016), thishas rarely been tested. Captive individuals mayexhibit different behaviors and/or kinematics thanwild counterparts (McPhee & Carlstead 2010), whichcould potentially influence accelerometer signatures.Because polar bears use both sea ice and terrestrialhabitats and because differences in habitat substrateor gradient could also affect accelerometer signa-tures (Bidder et al. 2012, Shepard et al. 2013,McClune et al. 2014), we examined data from wildpolar bears in both of these habitats. Lastly, becausesampling frequency affects the longevity of accelero -meters during deployment as well as computationalpower for analyses, we evaluated the ability of acce -lero meters to predict wild polar bear behaviors using3 different sampling frequencies (16, 8, and 4 Hz).

MATERIALS AND METHODS

Accelerometer recordings on captive bears

We deployed collars with archival loggers (TDR10-X-340D; Wildlife Computers) on 3 adult female polarbears Ursus maritimus housed at the Alaska Zoo,Oregon Zoo, and San Diego Zoo, USA, as well as 2adult female brown bears U. arctos housed at theBear Research, Education, and Conservation Centerat Washington State University (WSU; Table 1), USA.Ar chi val loggers recorded tri-axial acceleration (ms−2) at 16 Hz (range: ±20 m s−2), time-of-day, and

20

Pagano et al.: Accelerometers identify polar bear behaviors

wet/dry conductions (via an on-board conductivitysensor; Fig. 1). Conductivity data were sampled at1 Hz. Bears at the Oregon and San Diego Zoos weretrained to voluntarily place their heads into crates inwhich collars could be applied or removed, and worecollars for 1 to 4 h sessions. Bears at the Alaska Zooand WSU were anesthetized for collaring, with acombination of tiletamine HCl and zolazepam HCl(Telazol®; Pfizer Animal Health) and dexmedeto -midine HCl (Dexdomitor®; Pfizer Animal Health)(Teisberg et al. 2014). Following collar placement, theeffect of the anesthetic were reversed with atipame-zole HCl (Antisedan®; Pfizer Animal Health). Weused release mechanisms (Lotek Wireless) to removecollars from bears at the Alaska Zoo and WSU. We

matched accelerometer recordings tothe behaviors of captive bears whilethey moved freely around enclosuresbased on visual examination of time-stamped video recordings (Sony cam-corder model DCR-TRV280 or OpenEyeDigital Video Security Solutions).

Accelerometer recordings onfree-ranging polar bears

We deployed GPS-equipped videocamera collars (Exeye) and archival log-gers (TDR10-X-340D; Wild life Comput-ers) on 4 adult female polar bears and 1subadult female polar bear captured onthe sea ice of the southern Beaufort Sea

in April 2014 and 2015 (hereafter ‘ice bears’) and 2subadult polar bears (1 male and 1 female) capturedon land on Akimiski Island, Nunavut, Canada, inSeptember 2015 (hereafter ‘land bears’; Table 1).Video collars, including archival loggers and releasemechanisms, weighed 1.6 to 2.1 kg (0.8 to 1.5% ofbody mass of bears in this study). We captured polarbears by injecting them with immobilizing drugsthrough projectile syringes fired from a helicopter.On the sea ice, we anesthe tized bears using a combi-nation of tiletamine HCl and zolazepam HCl (Tela-zol®) with no reversal (Stirling et al. 1989). On land,we anesthetized bears with a combination of medeto-midine (Domitor®; Pfizer Animal Health) and tileta-mine HCl and zolazepam HCl (Telazol®) andreversed with atipamezole HCl (Antisedan®) (Cattetet al. 1997). Archival loggers were attached to collarsin the same location and orientation as captivedeployments (Fig. 1) and similarly recorded tri-axialacceleration at 16 Hz (range: ±20 m s−2), time-of-day,and wet/dry conductions (via an on-board conductiv-ity sensor). Conductivity data were sampled at 1 Hz.Video cameras were programmed to record at vary-ing frequencies during daylight periods (seeTable S1 in the Supplement at www.int-res.com/articles/ suppl/ n032 p019 _ supp. pdf) and programmedto turn off if the temperature of the collar fell below−17°C to protect video equipment. Collars deployedon ice and land bears were recovered 4 to 23 d fol-lowing deployment, either by recapture of the indi-vidual or by remote activation of the collar releaseand retrieval of the dropped collar by the field crew.We matched accelerometer data to behavior of iceand land bears based on visual examination of thetime-stamped video recordings from the collar.

21

Species Sex Age class Body Locationmass (kg)

Polar bear Female Adult 288 Alaska ZooPolar bear Female Adult 212 Oregon ZooPolar bear Female Adult 237 San Diego ZooBrown bear Female Adult 126 Washington State UniversityBrown bear Female Adult 126 Washington State UniversityPolar bear Female Adult 173 Southern Beaufort SeaPolar bear Female Adult 176 Southern Beaufort SeaPolar bear Female Adult 199 Southern Beaufort SeaPolar bear Female Adult 172 Southern Beaufort SeaPolar bear Female Subadult 141 Southern Beaufort SeaPolar bear Male Subadult 186 Akimiski IslandPolar bear Female Subadult 140 Akimiski Island

Table 1. Polar bears Ursus maritimus and brown bears U. arctos wearing col-lars with tri-axial accelerometers that were video recorded (captive bears) or

that wore video-equipped collars (wild bears)

Fig. 1. Orientation of an archival logger containing a tri- axial accelerometer attached to a collar for use on polar

Ursus maritimus and brown bears U. arctos

http://www.int-res.com/articles/suppl/n032p019_supp.pdfhttp://www.int-res.com/articles/suppl/n032p019_supp.pdf

Endang Species Res 32: 19–33, 2017

Behaviors

Behaviors were annotated based on the video dataon a per second basis. For bears that were anesthe -tized, we excluded behaviors on the day of captureand during retrieval of the collar. Resting behaviorsincluded standing, sitting, and lying down. Headmovements while standing, sitting, or lying downwere included as resting behaviors, but limb move-ments were treated as transitionary behaviors (Knud-sen 1978, Williams 1983). Swimming included surface swimming and diving. We excluded fromana lyses any behaviors that were not indicative ofnatural movements in captive bears (e.g. stereotypicbehaviors), were rare (e.g. fighting, breeding, drink-ing), were transitionary, or were non-descript.

Modeling

We derived summary statistics from the accelerom-eter data and linked the accelerometer data with cor-responding behaviors of interest (SAS version 9.3;SAS Institute). We converted accelerometer meas-ures from m s−2 to g (1 g = 9.81 m s−2). We calculatedmagnitude (Q) as a fourth dimension, where

(Nathan et al. 2012). We used a 2 s running mean ofthe raw acceleration data to calculate static accelera-

tion (gravitational acceleration) and subtracted thestatic acceleration from the raw acceleration data tocalculate dynamic acceleration (Wilson et al. 2006,Shepard et al. 2008). We calculated overall dynamicbody acceleration (ODBA) as the absolute sum ofdynamic acceleration across the 3 axes (Wilson et al.2006). We used a Fast Fourier Transform to calculatethe dominant power spectrum (dps) and frequency(fdps) for each axis (Watanabe et al. 2005, Shamoun-Baranes et al. 2012). In total, we derived 25 predictorvariables based on previous accelerometer studies(e.g. Watanabe et al. 2005, Nathan et al. 2012,Shamoun-Baranes et al. 2012, Wang et al. 2015). Pre-dictor variables were extracted from the accelerome-ter data over 2 s intervals; mean conductivity data(wet/dry) was also extracted over 2 s intervals usingprogram R (R Core Team 2014) (Table 2). Video-linked behaviors that lasted less than 2 s wereexcluded from analyses. We used a random forestsupervised machine learning algorithm (Breiman2001) in R (‘RandomForest’ package) to predict polarbear behaviors. Random forest models use multipleclassification trees from a random subset of predictorvariables and then replicate this process over multi-ple iterations using a subset of the data for each iter-ation to determine the best variables for making pre-dictions (Breiman 2001). An estimate of error isderived by using the remaining data not used in eachiteration to test the predictive ability of the model,which is termed the ‘out-of-bag’ (OOB) error rate

Q heave +surge +sway2 2 2=

22

Parameter Label Definition

Static acceleration (g) staticX, staticY, staticZ, staticQ Mean static acceleration along the surge, heave, sway, and magnitude axes

Maximum dynamic body mdbaX, mdbaY, mdbaZ, mdbaQ Maximum dynamic body acceleration along the acceleration (g) surge, heave, sway, and magnitude axes

Standard deviation dynamic stdbaX, stdbaY, stdbaZ, stdbaQ Standard deviation dynamic body acceleration body acceleration (g) along the surge, heave, sway, and magnitude axes

Overall dynamic body odbaX, odbaY, odbaZ, ODBA Mean dynamic acceleration body acceleration along acceleration (g) the surge, heave, and sway axes. ODBA = odbaX +

odbaY + odbaZ

Dominant power spectrum dpsX, dpsY, dpsZ, dpsQ Maximum power spectral density of dynamic accelera-(g2 Hz−1) tion along the surge, heave, sway, and magnitude axes

Frequency at the dominant fdpsX, fdpsY, fdpsZ, fdpsQ Frequency at the maximum power spectral density of power spectrum (Hz) dynamic acceleration along the surge, heave, sway,

and magnitude axes

Mean wet/dry wetdry Mean conductivity determination of whether the tag is wet or dry (range: 0−255)

Table 2. Parameters extracted from tri-axial accelerometer and conductivity data and used in random forest models to predict wild polar bear Ursus maritimus behaviors. Respective acceleration measures from the surge (X), heave (Y), sway (Z), and

magnitude (Q) axes

Pagano et al.: Accelerometers identify polar bear behaviors

(Breiman 2001, Liaw & Wiener 2002). The randomforest algorithm has previously shown high accuracy(>80%) for predicting animal behaviors fromaccelerometer data (Nathan et al. 2012, Resheff et al.2014, Graf et al. 2015, Lush et al. 2015, Rekvik 2015,Wang et al. 2015, Alvarenga et al. 2016). We fit 500classification trees to each training dataset and useda random subset of 5 predictor variables for each splitin the tree.

Analyses

Unbalanced datasets can bias the predictive abilityof classification algorithms toward the most dominantclasses (Chen et al. 2004). Therefore, we performed 3initial analyses to test the effect of uneven distribu-tions on predictive ability. The first analysis used anuneven distribution in which for ice and land bears,we randomly selected 70% of each behavior for thetraining dataset and used the remaining 30% to testthe predictive ability of the random forest algorithm(e.g. Nathan et al. 2012, Alvarenga et al. 2016). Forcaptive polar and brown bears we used the entiredatasets to train the random forest algorithm. Thesecond analysis used a subsampling approach inwhich we attempted to reduce the uneven distribu-tion of more frequent behaviors (e.g. resting) in ourtraining dataset. To reduce the uneven distribution ofbehaviors in the dataset from ice bears, we randomlyselected 5% of the resting behaviors, 30% of thewalking behaviors, and 70% of each of the remain-ing behaviors for training the random forest algo-rithm. We used the remaining data from ice bears fortesting predictions. To reduce the uneven distribu-tion of the dataset from land bears, we randomlyselected 5% of the resting behaviors and 70% ofeach of the remaining behaviors for training andused the remaining data to test predictions. To re -duce the uneven distribution of the datasets fromcaptive polar bears and brown bears, we randomlyselected 10% of the resting behaviors, 30% of thewalking behaviors, and 100% of each of the remain-ing behaviors for training the random forest algo-rithm. The third analysis used a completely balanceddistribution in which we used identical sample sizesof 500 observations for each behavior in the trainingdataset and the remaining observations to test andexcluded behaviors with less than 500 observations.Based on these 3 analyses, we used the sampling distribution (i.e. uneven, subsampled, or balanceddistribution) with the greatest predictive ability forfurther analyses.

To evaluate our ability to predict behaviors of icebears, we used 3 different datasets to train the ran-dom forest models and evaluated the ability of eachof these models. First, we used a random subset ofthe data from ice bears as the training dataset andthe remaining data from ice bears to test predictions(testing dataset). Second, we used the data from cap-tive polar bears as the training dataset. Third, weused the data from captive brown bears as the train-ing dataset.

To evaluate our ability to predict behaviors of landbears, we conducted 4 additional analyses. First, weused a random subset of the data from land bears asthe training dataset and the remaining data fromland bears to test predictions (testing dataset). Sec-ond, we used the training data from ice bears as thetraining dataset. Third, we used the training datafrom captive polar bears as the training dataset.Fourth, we used the training data from captive brownbears as the training dataset.

To examine the effect of sampling frequency onour ability to discriminate behaviors, we subsampledour 16 Hz accelerometer data to lower data acquisi-tion rates of 8 and 4 Hz using SAS, and repeated thepredictive analyses above for both ice and landbears.

Predicted behaviors were categorized as true posi-tive (TP) if they correctly matched the actual behav-ior, true negative (TN) if they correctly identified as adifferent behavior, false positive (FP) if they incor-rectly identified the behavior, and false negative(FN) if they incorrectly identified as a different be -havior. We evaluated the predictive abilities of thesemodels based on Matthews’ correlation coefficient(MCC; e.g. Basu et al. 2013, Martins et al. 2016), thepercent precision, recall, and F-measure. We usedMCC in place of accuracy due to the unbalancednature of our dataset.

MCC, ,

provides a measure of the agreement between thepredicted and actual classifications, where +1 repre-sents a perfect prediction and −1 represents total dis-agreement (Matthews 1975). Precision is the propor-tion of positive classifications that were correctlyclassified (TP/TP + FP), recall is the probability that abehavior will be correctly classified (TP/TP + FN),and F-measure is the harmonic mean of precisionand recall (2 × precision × recall/precision + recall).We used 2 sample t-tests to evaluate whether MCC,precision, and recall differed significantly using a16 Hz sampling frequency compared to either an 8 or

TP TN–FP FN

TP+FP TP+FN TN+FP TN+FN

× ×( ) × ( ) × ( ) × ( )

23

Endang Species Res 32: 19–33, 2017

4 Hz sampling frequency based on the ice and landdatasets.

RESULTS

Behavior on the sea ice

Video collars on ice bears Ursus maritimus re -corded 14 to 55 h of video (x− = 38 h, SD = 17 h, n = 5).For predicting the behavior of ice bears, we collecteda total of 140 h of video-linked accelerometer datafrom ice bears, 37 h from captive polar bears, and72 h from captive brown bears U. arctos. We identi-fied 10 different behaviors from ice bears, with rest-ing, walking, and eating being the most prevalent(Table 3). Ice bears ate recently killed adult, sub -adult, or pup ringed seals Pusa hispida, seal carcas -ses, bowhead whale Balaena mysti cetus carcasses, orunidentifiable carcasses. Captive polar bears con-sumed fish, and captive brown bears ate dry omni-vore chow. Captive brown bears also grazed ongrass, which was excluded from analyses predictingbehaviors of ice bears, but was included as eating forpredicting behaviors of land bears.

Our models using an uneven distribution of behav-iors in which we used 70% of each behavior from icebears and all of the available data from captive polaror brown bears exhibited 5% greater predictive abil-ity overall compared to the subsampled distribution,and 7% greater predictive ability overall comparedto the balanced distribution based onF-measure (see Table S2 in the Sup-plement at www. int-res. com/ articles/suppl/ n032 p019 _ supp. pdf). In particu-lar, the data sets with an uneven distri-bution ex hibited greater ability to dis-criminate less frequent behaviors suchas swimming, eating, and running(Table S2). Therefore, we used thedatasets with uneven distributions forsubsequent analyses (Table 3).

Our model with training data fromice bears had an OOB error rate of2.0% and exhibited the greatest pre-dictive abilities for all 10 behaviors(Fig. 2) compared to all other modelstested. Our models with training datafrom captive polar bears and brownbears had OOB error rates of 3.7 and0.5% respectively, indicating that bothmodels performed well in discriminat-ing captive behaviors. Both the ice

bear and captive polar bear models exhibited strongpredictive ability for identifying resting and walkingbehaviors in wild bears (>90% MCC, precision, re -call, and F-measure; Table 4 & Table S3 in the Sup-plement). Predictive abilities for other behaviors var-ied, with swimming and head shaking exhibitingstrong predictive ability using the ice bear model(>75% MCC, precision, recall, and F-measure), butlower predictive ability for eating, running, pounc-ing, grooming, digging, and rolling (Fig. 2, Tables 4& 5). The model from ice bears had particularlygreater ability than the captive polar bear model for

24

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Head shake

Roll

F-m

easu

re

Rest Walk Swim Eat Run GroomPounce Dig

Behavior

Ice bearsCaptive polar bears Captive brown bears

Fig. 2. Ability (F-measure) of the random forest model to predict 10 behaviorsof polar bears Ursus maritimus on the sea ice from 3 different trainingdatasets of accelerometer data. Ice bears: polar bears on the sea ice of the

southern Beaufort Sea

Behavior Captive Captive Wild Wildpolar brown ice landbears bears bears bears

Rest 53656 104838 163301 43132Walk 8962 33059 36083 958Swim 423 0 703 0Eat 2108 973 2100 1529Run 0 0 943 0Pounce 458 0 49 0Groom 3729 432 1138 289Dig 405 0 1194 0Head shake 86 14 435 19Roll 87 0 1473 0

Table 3. Number of 2 s long behaviors used in random foresttraining datasets for predicting behaviors of wild polar bearsUrsus maritimus. Ice bears: polar bears on the sea ice of thesouthern Beaufort Sea. Land bears: polar bears on Akimiski

Island, Nunavut

http://www.int-res.com/articles/suppl/n032p019_supp.pdfhttp://www.int-res.com/articles/suppl/n032p019_supp.pdf

Pagano et al.: Accelerometers identify polar bear behaviors

swimming, pouncing, and digging (Fig. 2, Table S3).The captive brown bear model provided weaker abil-ity to distinguish behaviors of ice bears for walking,eating, and grooming (

Endang Species Res 32: 19–33, 2017

lar mean ODBA values, but had differing mean staticacceleration values (Table S5). Eating and groominghad low values of static acceleration in the heave di-rection (staticY), which was indicative of a head-downposture. Walking and running exhibited periodic un-dulating patterns in static acceleration in the heavedirection (staticY; Fig. 4 & Fig. S1 in the Supplement),which was indicative of the bear’s head moving upand down as it stepped. Wet/dry conductivity whileswimming was lower for wild polar bears (x– = 81.9, SD= 81.5) than captive polar bears (x– = 205.3, SD = 57.8)and lower than all other behaviors (all x– > 234). A posthoc test excluding the conductivity variable reducedthe ability of the algorithm to correctly identify swim-ming be haviors using the training data set for icebears (MCC = 0.47, pre cision = 0.77, recall = 0.29, F-measure = 0.42) with a high rate of swimming behav-iors misclassified as resting.

Behaviors on land

Video collars on land bears recorded 19 to 36 h ofvideo (x– = 27 h, SD = 12 h, n = 2) and in total we col-lected 36 h of video-linked accelero meter data for thebehaviors of interest. We identified 5 different be ha -viors from land bears, with resting being the mostprevalent, followed by eating (Table 3). Eating onland consisted of berries, primarily crowberries Em -petrum nigrum.

Our model with training data from land bears hadan OOB error rate of 0.5% and had the greatest success in discriminating on-land behaviors (Fig. 5,Table S6). All behaviors, except for grooming andhead shaking, had MCC, precision, recall, and F-measure values >90% using the model from landbears (Fig. 5, Table S6). In particular, the model fromland bears was able to distinguish eating (MCC =0.95, precision = 0.95, recall = 0.96, F-measure =0.95), which was not possible with the other datasets.Our model with training data from ice bears had suc-cess in discriminating resting behaviors on land(MCC = 0.60, precision = 0.96, recall = 1.0, F-mea-sure = 0.98) and walking on land (MCC = 0.82, preci-sion = 0.89, recall = 0.76, F-measure = 0.82), but eat-ing was often misclassified as resting or walking (FP).The captive polar bear model performed similarly tothe model from ice bears for discriminating behaviorson land (Fig. 5). The captive brown bear model per-formed less well than the other models for discrimi-nating walking on land, but otherwise performedsimilarly to the models from ice bears and captivepolar bears (Fig. 5).

Sampling frequency

The OOB error rate using the data from ice bears in-creased from 2.0 to 2.2% using an 8 Hz sampling fre-quency and to 2.6% using a 4 Hz sampling frequency.OOB error rate using data from land bears increasedfrom 0.5 to 0.6% at 8 Hz and to 0.8% at 4 Hz. Predic-tive ability using an 8 Hz sampling frequency wasnearly identical to 16 Hz among all behaviors usingthe dataset from ice bears (t58 = 0.70, p = 0.24) andland bears (t28 = 0.61, p = 0.27) based on MCC, preci-sion, and recall. Predictive ability using a 4 Hz sam-pling frequency was lower than predictive ability us-ing 16 Hz for ice bears (t55 = 1.8, p = 0.04), but not forland bears (t27 = 0.59, p = 0.28). In particular, theability to discriminate the high intensity behaviors ofpouncing and head shaking declined using a 4 Hzsampling rate (Fig. 6).

DISCUSSION

Our results show that tri-axial accelerometers incombination with measures of conductivity can reli-ably distinguish the 3 most common behaviors ofwild polar bears Ursus maritimus (resting, walking,and swimming; Stirling 1974, Latour 1981, Hansson& Thomassen 1983, Lunn & Stirling 1985). This willprovide a method to remotely document the activitybudgets of these far-ranging animals, which can befurther linked with location data from satellite collarsto examine the effects of habitat on behavior andenergy expenditure. Our results indicate that differ-ences among habitats and species can impact theability to discriminate behaviors in wild individualsusing accelerometers. We found no loss in predictiveability using an 8 Hz sampling frequency, whichwould allow for twice the battery longevity of a 16 Hzrate and reduce the computational power needed foranalyses. Although accelerometer studies on smallerspecies appear to require greater sampling frequen-cies (e.g. >30 Hz; Broell et al. 2013, Brown et al.2013), our results are similar to data obtained byRekvik (2015) from captive brown bears U. arctos,and by Wang et al. (2015) from captive mountainlions Puma concolor, which both found little loss inpredictive ability at sampling frequencies ≥8 Hz.

Habitat effects

Our results indicate that accelerometer signatureson sea ice are similar to signatures on land for most

26

Pagano et al.: Accelerometers identify polar bear behaviors 27

Fig

. 4.

Acc

eler

omet

er s

ign

atu

res

of s

tati

c ac

cele

rati

on i

n t

he

surg

e (X

), h

eave

(Y

), a

nd

sw

ay (

Z)

dir

ecti

ons

and

ove

rall

dyn

amic

acc

eler

atio

n (

OD

BA

) w

hil

e w

alk

ing

, sw

imm

ing

, sta

nd

ing

, an

d e

atin

g a

sea

l fro

m a

n a

du

lt f

emal

e p

olar

bea

r U

rsu

s m

arit

imu

son

th

e se

a ic

e of

th

e so

uth

ern

Bea

ufo

rt S

ea

Endang Species Res 32: 19–33, 2017

behaviors, but eating berries by land bears had a dis-tinct signature that our ice bear model and captivebear models misclassified as grooming, resting, orwalking. This highlights the value in linking obser-vational and acce lero meter data from wild subjectsover multiple time periods and habitats, and theimportance of accounting for as many behaviors aspossible in training datasets. Knowledge of eatingfrequency and du ration would provide insight in de -termining foraging success, an im portant determi-

nant of individual re productive successand survival (Stirling et al. 1999, Regehr etal. 2007, 2010). Al though we had successdiscriminating eating events by land bears,we had lower precision and recall in dis-criminating eating events by ice bears. Thiswas likely related in part to the movementpattern of bears eating berries, in whichthey typically stood with their head downand grazed. Conversely, bears eating onthe sea ice exhibited a variety of positionsincluding standing, sitting, and lyingdown, and both tore pieces of food fromseals or gnawed on carcasses. Since mostkill events involve bears pouncing on theirseal prey (Stirling 1988, Derocher 2012),we may be able to identify successful killsbased on the combination of a pouncingsignature followed by eating signatures(e.g. Williams et al. 2014), but this requiresfurther evaluation. Additionally, feeding ona seal would typically last for a prolonged

period; hence, if the model primarily predicted eat-ing over a prolonged period this could be used as anindication of a feeding event, but this also requiresfurther evaluation.

Use of captive animals and surrogate species

Our ability to discriminate behaviors was greatlyimproved by including data from free-ranging polar

28

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Roll

F-m

easu

re

Rest Walk Swim Eat Run Pounce Groom Dig Headshake

Behavior

16 Hz

8 Hz

4 Hz

Fig. 6. Ability (F-measure) of a random forest model to predict behaviors of polar bears Ursus maritimus on the sea ice using 3 different accelerometer sampling frequencies

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Rest Walk Eat Groom Head shake

F-m

easu

re

Behavior

Land bearsIce bearsCaptive polar bears Captive brown bears

Fig. 5. Ability (F-measure) of a random forest model to predict behaviorsof polar bears Ursus maritimus on land from 4 different training datasets.Ice bears: polar bears on the the sea ice of the southern Beaufort Sea.

Land bears: polar bears on Akimiski Island, Nunavut

Pagano et al.: Accelerometers identify polar bear behaviors

bears rather than using data from captive bearsalone. However, resting and walking could be reli-ably discriminated using data from either captive orwild polar bears. This illustrates the value of collect-ing data from captive individuals when data collec-tion is difficult or impossible from wild counterparts.However, data from captive brown bears exhibitedpoorer performance for predicting active behaviorsin wild polar bears. This may be related to differ-ences in walking kinematics between polar andbrown bears as well as potential differences in limblengths between the species (Renous et al. 1998).Additionally, polar bears have longer necks relativeto their body size than other ursid species (DeMaster& Stirling 1981), which could also affect accelero -meter signatures from a neck-worn collar. AlthoughCampbell et al. (2013) proposed the use of surrogatespecies to predict the behaviors of other species, ourfindings suggest that polar bear accelerometer signa-tures are likely species- and habitat-specific, at leastfor distinguishing specific behaviors. The brown bearmodel did reliably distinguish resting behavior inwild polar bears, which suggests that surrogate spe-cies could be used to distinguish coarse activity pat-terns such as active versus inactive (e.g. Gervasi etal. 2006, Ware et al. 2015).

Our analyses indicate that conductivity measuresare needed to reliably discriminate swimming.Greater conductivity measures in captive polar bearsthat were swimming in fresh water likely caused thepoorer performance for discriminating swimming inwild polar bears that were swimming in salt water.For pouncing, captive polar bears pounced on largeplastic barrels, which resulted in similar measures ofODBA as wild counterparts, but had different signa-tures of static acceleration (i.e. body posture). Dig-ging by wild bears, which was often through snowand ice into subnivean lairs to locate seals, exhibitedgreater ODBA measures and slightly different staticacceleration than captive bears digging in snow andice. These results suggest that some behaviors ofcaptive bears may not fully reflect behaviors of theirwild counterparts, which further illustrates the valueof collecting simultaneous observational data (e.g. vi -deo) from free-ranging individuals to calibrate ac -celerometer-based behavioral data.

Accelerometer attachment

Regardless of which training dataset was used,we found lower precision and recall for predicting5 of the behaviors tested for bears on the sea ice.

Eating, grooming, and rolling had high rates ofmisclassifications as resting, whereas running anddigging had high rates of misclassifications aswalking. These re sults suggest the random forestalgorithm could be prone to slightly overestimatethe amount of true resting and walking behaviorsin quantifying activity budgets. Our lower precisionand recall for discriminating some behaviors waslikely due in part to the attachment of the accel -erometer on a collar. Al though a number of studieshave successfully discriminated behaviors usingaccelerometers on collars (Watanabe et al. 2005,Martiskainen et al. 2009, Soltis et al. 2012, McCluneet al. 2014, Lush et al. 2015, Rekvik 2015, Wang etal. 2015), many of these studies limited their analy-ses to 4 or 5 behaviors or documented high misclas-sification rates for distinguishing some behaviors.Wang et al. (2015) similarly reported low accuracyof accelerometers on collars for predicting eatingand grooming by captive mountain lions, and Lushet al. (2015) reported low ac curacy for predictingsome behaviors, including grooming, in wild brownhares Lepus europaeus. Attach ment of the accel -erometer to a collar, as opposed to attachmentdirectly on the animal, likely introduces noise inthe data due to independent collar motion (i.e. thecollar must be fitted to ensure animals do notremove it, but loose enough to accommodatepotential changes in body mass) and may reducethe ability of the accelerometer to detect some lowintensity movements (Shepard et al. 2008). Theeffect of independent collar motion is evident inour large values of ODBA when bears shook theirheads. This behavior may be useful for identifyingthe end of a swim, as bears are known to shakeand roll in the snow following a swim. Additionally,our ability to discriminate head shaking allows forexcluding it from potential energetic analyses usingaccelerometers. Use of a higher sampling frequencythan was used in this study (i.e. >16 Hz) couldpotentially improve the ability to discriminate somefine-scale body movements (Nathan et al. 2012)such as eating, though Wang et al. (2015) sampledat 64 Hz and had low accuracy in discriminatingeating behaviors of captive mountain lions.

Video calibration

Having video-linked observational data from cam-era-mounted collars on wild polar bears was the mostpractical method to calibrate accelerometers on free-ranging individuals. However, because the animal’s

29

Endang Species Res 32: 19–33, 2017

body was not visible in the video, some behaviorsmay have been incorrectly classified. For example,distinguishing walking versus running was oftenchallenging, as was determining when bears wereactively swimming versus resting in the water. Bothof these could have contributed to the misclassifica-tions between running and walking and swimmingand resting. Additionally, the models had greatersuccess discriminating behaviors as sample sizesincreased. Although unbalanced datasets are knownto affect the predictive ability of random forest algo-rithms (Chen et al. 2004), we found that the inclusionof larger sample sizes in the training dataset wasmore important than imbalance. This highlights thevalue of calibrating accelerometers from multipleindividuals over prolonged periods.

CONCLUSIONS

Our results underscore the importance of thor-oughly validating accelerometers for use in remotedetection of behavior, ideally on a species- and habi-tat-specific level. The use of tri-axial accelerometers,as shown here, will enable detailed assessments ofpolar bear behaviors to better understand polar bearhabitat use and the implications for energy demands.For example, measures of acceleration could be com-bined with measures of oxygen consumption fromcaptive bears while resting, walking, and swimmingto both quantify activity budgets and estimate theenergetic costs of these behaviors (e.g. Wilson et al.2006, 2012, Halsey et al. 2009, 2011, Gómez Laich etal. 2011, Williams et al. 2014). Future advances areneeded that would enable remote transmission ofraw accelerometer data to further enhance the appli-cability of these devices to animals occurring inremote environments and obviate the need for sensorrecovery. As declines in sea ice are expected to in -crease the activity rates of polar bears across much oftheir range (Derocher et al. 2004, Molnár et al. 2010,Sahanatien & Derocher 2012), the use of accelero -meters provides a method to monitor the impacts ofhabitat change on activity and energy budgets tobetter understand the implications for body condi-tion, reproductive success, and survival of this Arcticapex predator.

Acknowledgements. Procedures were approved by the Ani-mal Care and Use Committees of the University of Califor-nia, Santa Cruz; Washington State University; York Univer-sity; Oregon Zoo; Alaska Zoo; San Diego Zoo; OntarioMinistry of Natural Resources and Forestry; and the US

Geological Survey, Alaska Science Center. Research in theUSA was permitted under US Fish and Wildlife Service per-mits MA690038 and MA95406A. Research on AkimiskiIsland, Nunavut was approved under Nunavut WildlifeResearch Permit WL 2015-073. This work was supported byUS Geological Survey’s Changing Arctic Ecosystems Initia-tive. Additional support was provided by Polar Bears Inter-national; World Wildlife Fund (Canada); Ontario Ministry ofNatural Resources and Forestry; Natural Sciences and Engi-neering Research Council of Canada; Born Free Foundation;Helen McCrea Peacock Foundation; Institute for Conserva-tion Research, San Diego Zoo Global; and the InternationalAssociation for Bear Research and Management. Support forT.M.W. was provided by the National Science Foundation’sInstrument Development for Biological Research program.We thank Mehdi Bakhtiari (Exeye) for developing the videocollars used in this study. We thank the bear keeper andtrainer teams at the Oregon Zoo, San Diego Zoo, and AlaskaZoo for enabling data collection at their facilities. We thankStephen Atkinson, Tyrone Donnelly, Katie Florko, SarahHagey, Tim Moody, Kristin Simac, and Maria Spriggs forassistance in the field. We thank helicopter pilots Frank Ross(Soloy Helicopters) and Doug Holtby (OMNRF) for field sup-port. B. Battaile and 3 anonymous reviewers provided valu-able input on earlier versions of the manuscript. This re -search used resources of the Core Science Analytics andSynthesis Applied Research Computing program at the USGeological Survey. Use of trade names is for descriptive pur-poses only and does not imply endorsement by the US Gov-ernment. This paper was reviewed and approved by theUSGS under its Fundamental Science Practices policy(www.usgs.gov/fsp).

LITERATURE CITED

Alvarenga FAP, Borges I, Palkovi L, Rodina J, Oddy VH,Dobos RC (2016) Using a three-axis accelerometer toidentify and classify sheep behaviour at pasture. ApplAnim Behav Sci 181: 91−99

Amstrup SC, Marcot BG, Douglas DC (2008) A Bayesiannetwork modeling approach to forecasting the 21st cen-tury worldwide status of polar bears. In: DeWeaver ET,Blitz CM, Tremblay LB (eds) Arctic sea ice decline: observations, projections, mechanisms, and implications.Geophysics Monograph Series 180. American Geophysi-cal Union, Washington, DC, p 213−268

Atwood TC, Marcot BG, Douglas DC, Amstrup SC, RodeKD, Durner GM, Bromaghin JF (2016) Forecasting therelative influence of environmental and anthropogenicstressors on polar bears. Ecosphere 7: e01370

Basu I, Graupe D, Tuninetti D, Shukla P, Slavin KV, MetmanLV, Corcos DM (2013) Pathological tremor predictionusing surface EMG and acceleration: potential use in‘ON-OFF’ demand driven deep brain stimulator design.J Neural Eng 10: 036019

Bidder OR, Qasem LA, Wilson RP (2012) On higher ground: How well can dynamic body acceleration determinespeed in variable terrain? PLOS ONE 7: e50556

Bograd SJ, Block BA, Costa DP, Godley BJ (2010) Biolog-ging technologies: new tools for conservation. Introduc-tion. Endang Species Res 10: 1−7

Breiman L (2001) Random forests. Mach Learn 45: 5−32Broell F, Noda T, Wright S, Domenici P, Steffensen JF,

Auclair JP, Taggart CT (2013) Accelerometer tags: de -

30

https://doi.org/10.1016/j.applanim.2016.05.026https://doi.org/10.1002/ecs2.1370https://doi.org/10.1088/1741-2560/10/3/036019https://doi.org/10.1371/journal.pone.0050556https://doi.org/10.3354/esr00269https://doi.org/10.1023/A%3A1010933404324https://doi.org/10.1242/jeb.077396

Pagano et al.: Accelerometers identify polar bear behaviors

tecting and identifying activities in fish and the effect ofsampling frequency. J Exp Biol 216: 1255−1264

Bromaghin JF, McDonald TL, Stirling I, Derocher AE andothers (2015) Polar bear population dynamics in thesouthern Beaufort Sea during a period of sea ice decline.Ecol Appl 25: 634−651

Brown DD, Kays R, Wikelski M, Wilson R, Klimley AP (2013)Observing the unwatchable through acceleration log-ging of animal behavior. Anim Biotelem 1: 20

Campbell HA, Gao L, Bidder OR, Hunter J, Franklin CE(2013) Creating a behavioural classification module foracceleration data: using captive surrogate for difficult toobserve species. J Exp Biol 216: 4501−4506

Caro T (1999) The behavior-conservation interface. TrendsEcol Evol 14: 366−369

Cattet MRL, Caulkett NA, Polischuk SC, Ramsay MA (1997)Reversible immobilization of free-ranging polar bearswith medetomidine-zolazepam-tiletamine and atipame-zole. J Wildl Dis 33: 611−617

Chen C, Liaw A, Breiman L (2004) Using random forest tolearn imbalanced data. University of California, Berke-ley, CA. http: //statistics.berkeley. edu/ sites/ default/ files/tech- reports/666.pdf (accessed Oct 2015)

Cooke SJ (2008) Biotelemetry and biologging in endan-gered species research and animal conservation: rele-vance to regional, national, and IUCN Red List threatassessments. Endang Species Res 4: 165−185

Cooke SJ, Blumstein DT, Buchholz R, Caro T and others(2014) Physiology, behavior, and conservation. PhysiolBiochem Zool 87: 1−14

Costa DP, Huckstadt LA, Crocker DE, McDonald BI, GoebelME, Fedak MA (2010) Approaches to studying climatechange and its role on the habitat selection of Antarcticpinnipeds. Integr Comp Biol 50: 1018−1030

DeMaster DP, Stirling I (1981) Ursus maritimus. Mamm Spe-cies 145: 1−7

Derocher AE (2012) Polar bears: a complete guide to theirbiology and behavior. Johns Hopkins University Press,Baltimore, MD

Derocher AE, Lunn NJ, Stirling I (2004) Polar bears in awarming climate. Integr Comp Biol 44: 163−176

Durner GM, Douglas DC, Nielson RM, Amstrup SC and oth-ers (2009) Predicting 21st-century polar bear habitat dis-tribution from global climate models. Ecol Monogr 79: 25−58

Ferguson SH, Taylor MK, Messier F (2000) Influence of seaice dynamics on habitat selection by polar bears. Ecology81: 761−772

Gervasi V, Brunberg S, Swenson JE (2006) An individual-based method to measure animal activity levels: a test onbrown bears. Wildl Soc Bull 34: 1314−1319

Gómez Laich A, Wilson RP, Gleiss AC, Shepard ELC, Quin-tana F (2011) Use of overall dynamic body accelerationfor estimating energy expenditure in cormorants: Doeslocomotion in different media affect relationships? J ExpMar Biol Ecol 399: 151−155

Graf PM, Wilson RP, Qasem L, Hackländer K, Rosell F (2015)The use of acceleration to code for animal behaviours; acase study in free-ranging Eurasian beavers Castor fiber.PLOS ONE 10: e0136751

Halsey LG, Green JA, Wilson RP, Frappell PB (2009)Accelerometry to estimate energy expenditure duringactivity: best practice with data loggers. Physiol BiochemZool 82: 396−404

Halsey LG, Shepard ELC, Wilson RP (2011) Assessing the

development and application of the accelerometry tech-nique for estimating energy expenditure. Comp BiochemPhysiol A Mol Integr Physiol 158: 305−314

Hammond TT, Springthorpe D, Walsh RE, Berg-KirkpatrickT (2016) Using accelerometers to remotely and automat-ically characterize behavior in small animals. J Exp Biol219: 1618−1624

Hansson R, Thomassen J (1983) Behavior of polar bears withcubs in the denning area. Int Conf Bear Res Manage 5: 246−254

Hunter CM, Caswell H, Runge MC, Regehr EV, AmstrupSC, Stirling I (2010) Climate change threatens polar bearpopulations: a stochastic demographic analysis. Ecology91: 2883−2897

Kalxdorff SB (1997) Collection of local knowledge regard-ing polar bear habitat use in Alaska. US Fish andWildlife Service Tech Rep MMM 87-2. Marine Mam-mals Management, US Fish and Wildlife Service,Anchorage, AK

Knudsen B (1978) Time budgets of polar bears (Ursus mar-itimus) on North Twin Island, James Bay, during sum-mer. Can J Zool 56: 1627−1628

Kochnev AA, Etylin VM, Kavry VI, Siv-Siv EB, Tanko IV(2003) Traditional knowledge of Chukotka Native peo-ples regarding polar bear habitat use. Final Report. Pre-pared for the US National Park Service, Chukotka Asso-ciation of Traditional Marine Mammal Hunters, AlaskaNanuuq Commission, and Pacific Fisheries ResearchCenter (Chukotka Branch), Anchorage, AK

Latour PB (1981) Spatial relationships and behavior of polarbears (Ursus maritimus Phipps) concentrated on landduring the ice-free season of Hudson Bay. Can J Zool 59: 1763−1774

Liaw A, Wiener M (2002) Classification and regression byrandomForest. R News 2/3: 18−22

Lunn NJ, Stirling I (1985) The significance of supplementalfood to polar bears during the ice-free period of HudsonBay. Can J Zool 63: 2291−2297

Lush L, Ellwood S, Markham A, Ward AI, Wheeler P (2015)Use of tri-axial accelerometers to assess terrestrial mam-mal behavior in the wild. J Zool (Lond) 298: 257−265

Martins M, Santos C, Costa L, Frizera A (2016) Featurereduction and multi-classification of different assistivedevices according to the gait pattern. Disabil RehabilAssist Technol 11: 202−218

Martiskainen P, Järvinen M, Skön JP, Tiirikainen J, Koleh -mainen M, Mononen J (2009) Cow behaviour patternrecognition using three-dimensional accelerometer andsupport vector machines. Appl Anim Behav Sci 119: 32−38

Matthews BW (1975) Comparison of the predicted andobserved secondary structure of T4 phage lysozyme.Biochim Biophys Acta 405: 442−451

Mauritzen M, Belikov SE, Boltunov AN, Derocher AE andothers (2003) Functional response in polar bear habitatselection. Oikos 100: 112−124

McClune DW, Marks NJ, Wilson RP, Houghton JDR andothers (2014) Tri-axial accelerometers quantify be -haviour in the Eurasian badger (Meles meles): towardsan automated interpretation of field data. Anim Biotelem2: 5

McClune DW, Marks NJ, Delahay RJ, Montgomery WI,Scantlebury DM (2015) Behavior-time budget and func-tional habitat use of free-ranging European badger(Meles meles). Anim Biotelem 3: 7

31

https://doi.org/10.1890/14-1129.1https://doi.org/10.1186/2050-3385-1-20https://doi.org/10.1242/jeb.089805https://doi.org/10.1016/S0169-5347(99)01663-8https://doi.org/10.7589/0090-3558-33.3.611https://doi.org/10.3354/esr00063https://doi.org/10.1086/671165https://doi.org/10.1093/icb/icq054https://doi.org/10.2307/3503828https://doi.org/10.1093/icb/44.2.163https://doi.org/10.1890/07-2089.1https://doi.org/10.1890/0012-9658(2000)081%5b0761%3AIOSIDO%5d2.0.CO%3B2https://doi.org/10.2193/0091-7648(2006)34%5b1314%3AAIMTMA%5d2.0.CO%3B2https://doi.org/10.1016/j.jembe.2011.01.008https://doi.org/10.1371/journal.pone.0136751https://doi.org/10.1086/589815https://doi.org/10.1016/j.cbpa.2010.09.002https://doi.org/10.1242/jeb.136135https://doi.org/10.1890/09-1641.1https://doi.org/10.1139/z78-224https://doi.org/10.1139/z81-242https://doi.org/10.1139/z85-340https://doi.org/10.1111/jzo.12308https://doi.org/10.3109/17483107.2015.1079652https://doi.org/10.1016/j.applanim.2009.03.005https://doi.org/10.1016/0005-2795(75)90109-9https://doi.org/10.1034/j.1600-0706.2003.12056.xhttps://doi.org/10.1186/2050-3385-2-5https://doi.org/10.1186/s40317-015-0025-z

Endang Species Res 32: 19–33, 2017

McPhee ME, Carlstead K (2010) The importance of main-taining natural behaviors in captive mammals. In: Kleiman DG, Thompson KV, Baer CK (eds) Wild mam-mals in captivity: principles and techniques for zoo man-agement, 2nd edn. University of Chicago Press, Chicago,IL, p 303−313

Molnár PK, Derocher AE, Thiemann GW, Lewis MA (2010)Predicting survival, reproduction and abundance ofpolar bears under climate change. Biol Conserv 143: 1612−1622

Nakamura I, Goto Y, Sato K (2015) Ocean sunfish rewarm atthe surface after deep excursions to forage for siphono -phores. J Anim Ecol 84: 590−603

Nathan R, Spiegel O, Fortmann-Roe S, Harel R, Wikelski M,Getz WM (2012) Using tri-axial acceleration data to iden-tify behavioral modes of free-ranging animals: generalconcepts and tools illustrated for griffon vultures. J ExpBiol 215: 986−996

Nelson RK (1966) Alaska Eskimo exploitation of the sea iceenvironment. Arctic Aeromedical Laboratory, Fort Wain-wright, AK

Obbard ME, Cattet MRL, Howe EJ, Middel KR and others(2016) Trends in body condition in polar bears (Ursusmaritimus) from the Southern Hudson Bay subpopula-tion in relation to changes in sea ice. Arctic Sci 2: 15−32

R Core Team (2014) R: a language and environment for sta-tistical computing. R Foundation for Statistical Comput-ing, Vienna

Regehr EV, Lunn NJ, Amstrup SC, Stirling I (2007) Effects ofearlier sea ice breakup on survival and population size ofpolar bears in western Hudson Bay. J Wildl Manag 71: 2673−2683

Regehr EV, Hunter CM, Caswell H, Amstrup SC, Stirling I(2010) Survival and breeding of polar bears in the south-ern Beaufort Sea in relation to sea ice. J Anim Ecol 79: 117−127

Rekvik GES (2015) Identification of five brown bear (Ursusarctos) behaviors using tri-axial accelerometers. MScthesis, Telemark University College

Renous S, Gasc JP, Abourachid A (1998) Kinematic analysisof the locomotion of the polar bear (Ursus maritimus,Phipps, 1774) in natural and experimental conditions.Neth J Zool 48: 145−167

Resheff YS, Rotics S, Harel R, Spiegel O, Nathan R (2014)AcceleRater: a web application for supervised learningof behavioral modes from acceleration measurements.Mov Ecol 2: 27

Rode KD, Amstrup SC, Regehr EV (2010) Reduced body sizeand cub recruitment in polar bears associated with seaice decline. Ecol Appl 20: 768−782

Rode KD, Peacock E, Taylor M, Stirling I, Born EW, LaidreKL, Wiig Ø (2012) A tale of two polar bear populations: ice habitat, harvest, and body condition. Popul Ecol 54: 3−18

Ropert-Coudert Y, Wilson RP (2005) Trends and perspec-tives in animal-attached remote sensing. Front Ecol Env-iron 3: 437−444

Sahanatien V, Derocher AE (2012) Monitoring sea ice habi-tat fragmentation for polar bear conservation. Anim Con-serv 15: 397−406

Schweinsburg RE, Lee LJ (1982) Movement of four satellite-monitored polar bears in Lancaster Sound, NorthwestTerritories. Arctic 35: 504−511

Shamoun-Baranes J, Bom R, van Loon EE, Ens BJ, Ooster-beek K, Bouten W (2012) From sensor data to animal

behaviour: an oystercatcher example. PLOS ONE 7: e37997

Shepard ELC, Wilson RP, Quintana F, Gómez Laich A andothers (2008) Identification of animal movement patternsusing tri-axial accelerometry. Endang Species Res 10: 47−60

Shepard ELC, Wilson RP, Rees WG, Grundy E, LambertucciSA, Vosper SB (2013) Energy landscapes shape animalmovement ecology. Am Nat 182: 298−312

Soltis J, Wilson RP, Douglas-Hamilton I, Vollrath F, King LE,Savage A (2012) Accelerometers in collars identifybehavioral states in captive African elephants Loxodontaafricana. Endang Species Res 18: 255−263

Stirling I (1974) Midsummer observations on the behavior ofwild polar bears (Ursus maritimus). Can J Zool 52: 1191−1198

Stirling I (1988) Polar bears. University of Michigan Press,Ann Arbor, MI

Stirling I, Latour PB (1978) Comparative hunting abilities ofpolar bear cubs of different ages. Can J Zool 56: 1768−1772

Stirling I, Spencer C, Andriashek D (1989) Immobilization ofpolar bears (Ursus maritimus) with Telazol® in the Cana-dian Arctic. J Wildl Dis 25: 159−168

Stirling I, Lunn NJ, Iacozza J (1999) Long-term trends in thepopulation ecology of polar bears in western Hudson Bayin relation to climatic change. Arctic 52: 294−306

Stirling I, Spencer C, Andriashek D (2016) Behavior andactivity budgets of wild breeding polar bears (Ursusmari timus). Mar Mamm Sci 32: 13−37

Sutherland WJ (1998) The importance of behavioral studiesin conservation biology. Anim Behav 56: 801−809

Taylor MK (1986) Movements of Alaskan polar bears instru-mented with satellite transmitters. Int Conf Bear ResManage 6: 103−104

Teisberg JE, Farley SD, Lynne Nelson O, Hilderbrand GVand others (2014) Immobilization of grizzly bears (Ursusarctos) with dexmedetomidine, tiletamine, and zola -zepam. J Wildl Dis 50: 74−83

Volpov BL, Hoskins AJ, Battaile BC, Viviant M and others(2015) Identification of prey captures in Australian furseals (Arctocephalus pusillus doriferus) using head-mounted accelerometers: field validation with animal-borne video cameras. PLOS ONE 10: e0128789

Vongraven D, Aars J, Amstrup S, Atkinson SN and others(2012) A circumpolar monitoring framework for polarbears. Ursus 23: 1−66

Voorhees H, Sparks R, Huntington HP, Rode KD (2014) Tra-ditional knowledge about polar bears (Ursus maritimus)in northwestern Alaska. Arctic 67: 523−536

Wang Y, Nickel B, Rutishauser M, Bryce C, Williams TM,Elkaim G, Wilmers CC (2015) Movement, resting, andattack behaviors of wild pumas are revealed by tri-axialaccelerometer measurements. Mov Ecol 3: 2

Ware JV, Rode KD, Pagano AM, Bromaghin J and others(2015) Validation of mercury tip-switch and accelerome-ter activity sensors for identifying resting and activebehavior in bears. Ursus 26: 86−96

Watanabe YY, Takahashi A (2013) Linking animal-bornevideo to accelerometers reveals prey capture variability.Proc Natl Acad Sci USA 110: 2199−2204

Watanabe S, Izawa M, Kato A, Ropert-Coudert Y, Naito Y(2005) A new technique for monitoring the detailedbehaviour of terrestrial animals: a case study with thedomestic cat. Appl Anim Behav Sci 94: 117−131

32

https://doi.org/10.1016/j.biocon.2010.04.004https://doi.org/10.1111/1365-2656.12346https://doi.org/10.1242/jeb.058602https://doi.org/10.1139/as-2015-0027https://doi.org/10.2193/2006-180https://doi.org/10.1111/j.1365-2656.2009.01603.xhttps://doi.org/10.1163/156854298X00156https://doi.org/10.1186/s40462-014-0027-0https://doi.org/10.1890/08-1036.1https://doi.org/10.1007/s10144-011-0299-9https://doi.org/10.1890/1540-9295(2005)003%5b0437%3ATAPIAR%5d2.0.CO%3B2https://doi.org/10.1111/j.1469-1795.2012.00529.xhttps://doi.org/10.1371/journal.pone.0037997https://doi.org/10.3354/esr00084https://doi.org/10.1086/671257https://doi.org/10.3354/esr00452https://doi.org/10.1139/z74-157https://doi.org/10.1139/z78-242https://doi.org/10.7589/0090-3558-25.2.159https://doi.org/10.14430/arctic935https://doi.org/10.1111/mms.12291https://doi.org/10.1006/anbe.1998.0896https://doi.org/10.7589/2012-11-273https://doi.org/10.1371/journal.pone.0128789https://doi.org/10.14430/arctic4425https://doi.org/10.1186/s40462-015-0030-0https://doi.org/10.2192/URSUS-D-14-00031.1https://doi.org/10.1073/pnas.1216244110https://doi.org/10.1016/j.applanim.2005.01.010

Pagano et al.: Accelerometers identify polar bear behaviors

Williams TM (1983) Locomotion in the North Americanmink, a semi-aquatic mammal. I. Swimming energeticsand body drag. J Exp Biol 103: 155−168

Williams TM, Wolfe L, Davis T, Kendall T and others (2014)Instantaneous energetics of puma kills reveal advantageof felid sneak attacks. Science 346: 81−85

Wilmers CC, Nickel B, Bryce CM, Smith JA, Wheat RE,Yovovich V (2015) The golden age of bio-logging: howanimal-borne sensors are advancing the frontiers of eco -logy. Ecology 96: 1741−1753

Wilson RP, White CR, Quintana F, Halsey LG, Liebsch N,Martin GR, Butler PJ (2006) Moving towards accelerationfor estimates of activity-specific metabolic rate in free-living animals: the case of the cormorant. J Anim Ecol 75: 1081−1090

Wilson RP, Shepard ELC, Liebsch N (2008) Prying into the

intimate details of animal lives: use of a daily diary onanimals. Endang Species Res 4: 123−137

Wilson RP, Quintana F, Hobson VJ (2012) Construction ofenergy landscapes can clarify the movement and distri-bution of foraging animals. Proc R Soc B 279: 975−980

Wilson RR, Horne JS, Rode KD, Regehr EV, Durner GM(2014) Identifying polar bear resource selection patternsto inform offshore development in a dynamic and chang-ing Arctic. Ecosphere 5: 136

Yoda K, Sato K, Niizuma Y, Kurita M, Bost CA, Le Maho Y,Naito Y (1999) Precise monitoring of porpoising behav-iour of Adélie penguins determined using accelerationdata loggers. J Exp Biol 202: 3121−3126

Yoda K, Naito Y, Sato K, Takahashi A and others (2001) Anew technique for monitoring the behaviour of free-ranging Adélie penguins. J Exp Biol 204: 685−690

33

Editorial responsibility: Rory Wilson, Swansea, UK

Submitted: April 25, 2016; Accepted: October 13, 2016Proofs received from author(s): December 3, 2016

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=6854200&dopt=Abstracthttps://doi.org/10.1126/science.1254885https://doi.org/10.1890/14-1401.1https://doi.org/10.1111/j.1365-2656.2006.01127.xhttps://doi.org/10.3354/esr00064https://doi.org/10.1098/rspb.2011.1544https://doi.org/10.1890/ES14-00193.1http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=10539960&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11171350&dopt=Abstract