@article{ca7da418672a44d38a4bad1ceced1367,
title = "Coiban Agouti (Dasyprocta coibae) Density and Temporal Activity on Coiba Island, Veraguas, Panama",
abstract = "The Coiban agouti (Dasyprocta coibae) is a meso-sized rodent endemic to Coiba Island, Panama and is ostensibly classified as vulnerable to extinction. Despite this vulnerability, the ecology of the species remains unknown, and baseline estimates of population density and habitat use are needed to develop a scientifically-based conservation status and plan for the species. We recorded 2114 detections of Coiban agoutis using 30 camera traps broadly deployed across the range of habitat types on Coiba Island. We assessed 13 occupancy models using detections of unmarked agoutis, of which the best model estimated 73.3 agoutis/km2, which was positively related to proportion of primary forest. Coiban agouti detection was 86\% and was greater on government maintained trails than wildlife trails. Coiban agoutis were most active during daylight (06:00-19:00) hours and during late May to early June. Estimated Coiban agouti density was less than density estimates of related agouti species in tropical forests of Central America with greater predation risk, suggesting habitat may be limited for Coiban agoutis. Our results suggest that a camera trap survey would be effective for future population monitoring of Coiban agoutis, which could be designed to assess the population dynamics and guide the conservation of the species.",
keywords = "Camera trap, conservation, habitat, rodent, weather",
author = "Duquette, \{Jared F.\} and Luis Ure{\~n}a and Josu{\'e} Ortega and Iliana Cisneros and Ricardo Moreno and Flores, \{Eric E.\}",
note = "Occupancy modeling incorporating detection/non-detection of unmarked individuals provided an effective method to accurately and precisely estimate the density of Coiban agoutis. The precision of our agouti density estimate (0.2) was 11\% of the mean, which is satisfactory when compared to similar occupancy surveys of neotropical terrestrial mammals (e.g., Silver et al. 2004). Therefore, we suggest that our survey and modeling approach is suitable for estimating the density and abundance of Coiban agoutis, and likely other agouti species (Meyer et al. 2015). Agoutis have historically been surveyed using visual counts while walking along transects or mark-recapture studies, which could introduce greater bias from the shy nature of agoutis (Cant 1977), visual obstructions (e.g., vegetation), and difficulty of capturing agoutis (Smythe 1978). By deploying camera traps we reduced the likelihood of deterring agoutis due to human presence and non-invasively collected detection data over several months to relate to covariates of detection and abundance. We suggest our approach is more advantageous to estimate agouti density than visual transect count or mark-recapture surveys, especially considering the accuracy and precision of our density estimates. Future research is needed to develop a standardized camera trap survey for Coiban agoutis and assessment of additional covariates of detection (e.g., seasonality) and abundance (e.g., seed availability estimates). This information would be useful to understand the population dynamics and classify the conservation status of the Coiban agouti, particularly in the face of a changing climate that could considerably influence the Coiba Island landscape (Lyra et al. 2016) Acknowledgments: We thank H. Burnand, V. Bravo, Z. Capitan Barrios, R. Correa, D. Seiler, O. Shen, and L. V{\'a}squez for their dedication to setting and retrieving cameras. We thank J. Belant and F. Casta{\~n}eda for lending camera traps for the survey. We thank the Institute for Wildlife Studies for providing partial financial support for the primary author.",
year = "2017",
month = sep,
doi = "10.3106/041.042.0305",
language = "English (US)",
volume = "42",
pages = "153--160",
journal = "Mammal Study",
issn = "1343-4152",
publisher = "Mammalogical Society of Japan",
number = "3",
}