Scholarlink

Background <p>Two primary methods exist for estimating the daily travel distance by free-ranging animals: inference from animal location data recorded with biologging or wearable GNSS tags (e.g. GPS), and inference from snapshots of animal movement recorded by stationary sensors (e.g. camera traps). Formal comparisons between the two approaches, which differ in both their assumptions and in choices made during analyses, remain scarce.</p> Methods <p>Here, we compare GPS- and camera-based estimates of daily travel distance at matched temporal scales under both simulated and real-world scenarios. First, using simulated trajectories, we varied the GPS sampling interval and estimation method — continuous-time movement models (CTMM) vs. straight-line displacement (SLD) — as well as camera-analysis choices (activity correction, data sparsity) to test how well each approach recovers the true distance travelled. Second, we applied the same workflow to empirical datasets with overlapping GPS telemetry and camera-trap observations from the same populations.</p> Results <p>We found that, under the simulated movement and observation scenarios considered here, daily distance travelled could be estimated with limited bias from both camera-trap and GPS data using CTMM, but not with GPS data using SLD. Accurate estimation required correction for GPS-error, sufficiently high GPS sampling frequency, activity estimation for camera-trap data and adequate camera-trap data volume to estimate instantaneous speed and activity. Using empirical data from three case studies, we found that camera trapping and GPS-based CTMM estimates were of similar magnitude with overlapping confidence intervals, while GPS-based SLD estimates were lower than the others.</p> Conclusions <p>These findings suggest that, when appropriately implemented, both methods can yield similar and fairly unbiased estimates of daily travel distance. These results provide practical guidance for using and potentially combining GPS-based CTMM and camera trapping-based travel distances in density models such as REM (random encounter model), strengthening ecological inference and the effectiveness of wildlife monitoring and management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Camera traps and GPS telemetry can provide similar estimates of daily travel distance in terrestrial mammals

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月31日

查看全文
  • Thomas Osinga,
  • Henrik J. de Knegt,
  • Francesca Cagnacci,
  • Massimo Scandura,
  • Maik Henrich,
  • Marco Heurich,
  • Astrid Olejarz,
  • J. Marcus Rowcliffe,
  • Patrick A. Jansen

摘要

Background

Two primary methods exist for estimating the daily travel distance by free-ranging animals: inference from animal location data recorded with biologging or wearable GNSS tags (e.g. GPS), and inference from snapshots of animal movement recorded by stationary sensors (e.g. camera traps). Formal comparisons between the two approaches, which differ in both their assumptions and in choices made during analyses, remain scarce.

Methods

Here, we compare GPS- and camera-based estimates of daily travel distance at matched temporal scales under both simulated and real-world scenarios. First, using simulated trajectories, we varied the GPS sampling interval and estimation method — continuous-time movement models (CTMM) vs. straight-line displacement (SLD) — as well as camera-analysis choices (activity correction, data sparsity) to test how well each approach recovers the true distance travelled. Second, we applied the same workflow to empirical datasets with overlapping GPS telemetry and camera-trap observations from the same populations.

Results

We found that, under the simulated movement and observation scenarios considered here, daily distance travelled could be estimated with limited bias from both camera-trap and GPS data using CTMM, but not with GPS data using SLD. Accurate estimation required correction for GPS-error, sufficiently high GPS sampling frequency, activity estimation for camera-trap data and adequate camera-trap data volume to estimate instantaneous speed and activity. Using empirical data from three case studies, we found that camera trapping and GPS-based CTMM estimates were of similar magnitude with overlapping confidence intervals, while GPS-based SLD estimates were lower than the others.

Conclusions

These findings suggest that, when appropriately implemented, both methods can yield similar and fairly unbiased estimates of daily travel distance. These results provide practical guidance for using and potentially combining GPS-based CTMM and camera trapping-based travel distances in density models such as REM (random encounter model), strengthening ecological inference and the effectiveness of wildlife monitoring and management.