Geology Reports⌕ Search
Home/Archive/USGS

USGS · 70185240

Identifying and mitigating errors in satellite telemetry of polar bears

Stephen M. Arthur·Gerald W. Garner·Tamara L. Olson
View original source ↗Export metadata ↓

Abstract

Satellite radiotelemetry is a useful method of tracking movements of animals that travel long distances or inhabit remote areas. However, the logistical constraints that encourage the use of satellite telemetry also inhibit efforts to assess accuracy of the resulting data. To investigate effectiveness of methods that might be used to improve the reliability of these data, we compared 3 sets of criteria designed to select the most plausible locations of polar bears ( Ursus maritimus ) that were tracked using satellite radiotelemetry in the Bering, Chukchi, East Siberian, Laptev, and Kara seas during 1988-93. We also evaluated several indices of location accuracy. Our results suggested that, although indices could provide information useful in evaluating location accuracy, no index or set of criteria was sufficient to identify all the implausible locations. Thus, it was necessary to examine the data and make subjective decisions about which locations to accept or reject. However, by using a formal set of selection criteria, we simplified the task of evaluating locations and ensured that decisions were made consistently. This approach also enabled us to evaluate biases that may be introduced by the criteria used to identify location errors. For our study, the best set of selection criteria comprised: (1) rejecting locations for which the distance to the nearest other point from the same day was >50 km; (2) determining the highest accuracy code (NLOC) for a particular day and rejecting locations from that day with lesser values; and (3) from the remaining locations for each day, selecting the location closest to the location chosen for the previous transmission period. Although our selection criteria seemed unlikely to bias studies of habitat use or geographic distribution, basing selection decisions on distances between points might bias studies of movement rates or distances. It is unlikely that any set of criteria will be best for all situations; to make efficient use of data and minimize bias, these rules must be tailored to specific study objectives.

Explore related subjects

UrsusArticle

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Stephen M. Arthur, Gerald W. Garner, Tamara L. Olson. 1998. Identifying and mitigating errors in satellite telemetry of polar bears. https://pubs.usgs.gov/publication/70185240

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Detecting occurrence and timing of grizzly bear parturition based on anomalies in accelerometer data

Documenting natality of brown (grizzly) bears ( Ursus arctos ) is an important component of many population monitoring and research programs. Brown bears give birth during hibernation, so observation of litters is generally not feasible until females exit dens, and detection of litters can be compromised by poor observability. Using triaxial accelerometer activity data stored on radiocollars, we developed a technique that predicts births by the presence of brief upsurges in activity likely associated with postnatal maternal behaviors. We developed presence criteria with a training sample of known parturient females ( n = 22) to detect anomalies within time-series of daily motion-counts (i.e., activity readings >0 measured at 10-min intervals) during 25 December–7 March. To test performance, we applied criteria to a blinded sample of activity data obtained from female grizzly bears (captured and collared during 2012–2022) in 4 populations in interior North America ( n = 295). We assigned predicted status and compared assignments to reproductive status at first visual observation. The true positive rate was 91% ( n = 47 females observed with cubs of the year) and the false-positive rate was 17% ( n = 65 females observed with older offspring). Births were predicted for 50% of females with unknown reproductive status ( n = 48), 21% of females observed without offspring ( n = 114), and 10% of subadult females considered too young to reproduce ( n = 21). Dates of predicted births varied from 27 December to 28 February. Our anomaly detection technique was successful at estimating parturition events, and despite some error, indicated that a number of litters were born but not observed, presumably due to mortality in or shortly after den emergence. This technique provides an additional tool for supplementing visual observations for natality estimation and population modeling, provided that potential biases stemming from increased detection rates are considered.

Alaska, British Columbia, Idaho, Montana↗
USGS2025

Denning black bear response to anthropogenic disturbance and implications for cub survival in Florida

Wildlife research and management can be disruptive to wildlife. By advancing our understanding of the impacts of these activities, we can reduce adverse effects, improve decision-making, and enhance the outcomes of research and management. During 2017–2019, we observed the responses of denning female American black bears ( Ursus americanus ) to 3 types of routine research and management activities in Florida, USA: (1) a low-level, nonintrusive human approach near the natal den ( n H 44); (2) a high-level, intrusive human approach involving cub handling ( n H 42); and (3) a prescribed burn within 1 km of the den during the denning season ( n H 11). We measured responses (flight distance, time away, and postdisturbance denning behavior) using Global Positioning System collars programmed to record a location every 2 hours. We observed minimal response from bears to low-level human disturbances. In contrast, all bears fled after high-level human disturbances, with responses ranging from staying nearby throughout the disturbance and quickly returning to cubs, to fleeing several kilometers and abandoning cubs. On average, bears fled approximately 380 m from the den and returned to their cubs 7 hours postdisturbance. After returning, most bears relocated their cubs to a new den site, on average 125 m away. Responses to prescribed fire ranged from no measurable response and no den relocation to den site abandonment with cub mortality. Through generalized linear modeling, we found that adult female time away was positively associated with cub age. We found that annual cub survival was negatively associated with fire exposure in the den and with continued denning at a disturbed den site following high-level disturbance. In areas where bears are easily displaced from dens, these results provide insights that may improve bear research and habitat management decisions.

Florida↗
USGS2023

Enhancements to population monitoring of Yellowstone grizzly bears

In the Greater Yellowstone Ecosystem, counts of female grizzly bears ( Ursus arctos ) with cubs-of-the-year (females with cubs) from systematic aerial surveys and opportunistic ground sightings are combined with demographic data to derive annual population estimates. We addressed 2 limitations to the monitoring approach. As part of a rule set, a conservative distance of > 30 km currently is used as a threshold to assign sightings to unique females with cubs, resulting in underestimation bias. Using telemetry locations of females with cubs collected during 1997–2019, we created 1,000 data sets for each of 5 levels of simulated number of females with cubs, simulated sightings by selecting among these locations, and evaluated the classification performance of alternative distance criteria (12–30 km). Under all scenarios, 12–16-km criteria maximized classification performance and minimized estimation bias; the 16-km criterion was optimal for current conditions and sampling efforts. Our second objective was to test generalized additive models (GAMs) as a flexible trend analysis technique. We simulated 1,000 time series for each of 10 scenarios (10, 15, and 20% decline over periods of 5, 10, and 15 yrs, plus stability), applied GAMs, and assessed metrics associated with the posterior distribution of the instantaneous rate of change. We detected declines among > 99.6% of replicates under the 15 and 20% decline scenarios and in 84.7–94.7% of replicates under the 10% decline scenario. From decline onset to first detection, periods ranged from 3.7 (20% decline over 5 yrs) to 11.1 (10% decline over 15 yrs), with 3.9–8.8 years mean duration of detection events. The GAM approach allows detection of directional changes in population trend, including early warning metrics, and stabilization after such changes. Retrospective application of the 16-km distance criterion and GAMs resulted in higher population estimates and growth rates than are reported using current methods.

Wyoming↗