Variable terrestrial GPS telemetry detection rates: Addressing the probability of successful acquisitions. (27th April 2017)

• Record Type: Journal Article
• Authors: Ironside, Kirsten E.; Mattson, David J.; Choate, David; Stoner, David; Arundel, Terence; Hansen, Jered; Theimer, Tad; Holton, Brandon; Jansen, Brian; Sexton, Joseph O.; Longshore, Kathleen; Edwards, Thomas C.; Peters, Michael
• Is Part Of: Wildlife Society bulletin. Volume 41:Number 2(2017)
• Page Start: 329
• Page End: 341
• Publication Date: 2017-04-27
• Subjects: cougar -- desert bighorn sheep -- elk -- GPS telemetry -- location bias -- mountain lion -- mule deer; Wildlife management -- Periodicals; Wildlife conservation -- Periodicals; 333.9540973
• DOI: 10.1002/wsb.758
• Languages: English
• ISSNs: 0091-7648

Abstract

ABSTRACT: Studies using global positioning system (GPS) telemetry rarely result in 100% fix success rates (FSR), which may bias datasets because data loss is systematic rather than a random process. Previous spatially explicit models developed to correct for sampling bias have been limited to small study areas, a small range of data loss, or were study‐area specific. We modeled environmental effects on FSR from desert to alpine biomes, investigated the full range of potential data loss (0–100% FSR), and evaluated whether animal body position can contribute to lower FSR because of changes in antenna orientation based on GPS detection rates for 4 focal species: cougars ( Puma concolor ), desert bighorn sheep ( Ovis canadensis nelsoni ), Rocky Mountain elk ( Cervus elaphus nelsoni ), and mule deer ( Odocoileus hemionus ). Terrain exposure and height of over story vegetation were the most influential factors affecting FSR. Model evaluation showed a strong correlation (0.88) between observed and predicted FSR and no significant differences between predicted and observed FSRs using 2 independent validation datasets. We found that cougars and canyon‐dwelling bighorn sheep may select for environmental features that influence their detectability by GPS technology, mule deer may select against these features, and elk appear to be nonselective. We observed temporal patterns in missed fixes only for cougars. We provide a model for cougars, predicting fix success by time of day that is likely due to circadian changes in collar orientation and selection of daybed sites. We also provide a model predicting the probability of GPS fix acquisitions given environmental conditions, which had a strong relationship ( r 2 = 0.82) with deployed collar FSRs across species. Published 2017. This article is a U.S. Government work and is in the public domain in the USA. Abstract : This study investigates factors that contribute to missed GPS telemetry fixes and provides correction measures for detection bias. In the southwestern United States we found terrain, vegetation height, and collar orientation contributes to the probability of obtaining a fix.