How to monitor elusive lizards: comparison of capture-recapture methods on giant day geckos (Gekkonidae, Phelsuma madagascariensis grandis) in the Masoala rainforest exhibit, Zurich Zoo

Thomas Wanger, Iris Motzke, Samuel Furrer, Barry Brook, Bernd Gruber

    Research output: Contribution to journalArticle

    14 Citations (Scopus)

    Abstract

    Rapid and reliable estimation of population size is needed for the efficient monitoring of animal populations of conservation concern. Unfortunately, technical advances in this area have not been paralleled in uptake in conservation, which may be due to difficulties in implementation or the lack of general guidelines for application. Here we tested five different methods used to estimate population size [capture-mark-recapture (CMR), finite-mixture models, model averaging of finite-mixture models, accumulation curve methods (ACM), and the line transect method (LT)] using extensive captureâ¿¿recapture data of the giant day gecko (Gekkonidae, Phelsuma madagascariensis grandis, Gray 1870) at the Masoala rainforest exhibit, Zurich Zoo. When the complete data were analyzed [30 sessions (and 27 sessions for the LT)], all methods except the LT produced similar estimates of population size. The simple ACM gave a small coefficient of variation (CV), but did not cover the most likely value of population size at moderate sampling effort. Nevertheless, the ACM was the only method that showed a reasonable convergence when subsets of data were used. CMR and Pledger models included the reference value in their confidence intervals (CI) after 25 and 30 sessions, respectively. Although model averaging did slightly improve the estimate, the CV was still high for the full dataset. Our method of using subsets of data to test the robustness of estimates is simple to apply and could be adopted more widely in such analyzes to evaluate sensitivity to method of evaluation. In conclusion, simple accumulation methods showed similar efficiency to more complex statistical models, and are likely to be sufficiently precise for most conservation monitoring purposes.
    Original languageEnglish
    Pages (from-to)345-353
    Number of pages9
    JournalEcological Research
    Volume24
    DOIs
    Publication statusPublished - 2009

    Fingerprint

    zoo
    Gekkonidae
    zoos
    lizard
    rainforest
    rain forests
    lizards
    monitoring
    line transect
    population size
    methodology
    Phelsuma
    comparison
    method
    statistical models
    normal values
    confidence interval

    Cite this

    @article{861c63d187574b94bd5d9fc1f3bfbe55,
    title = "How to monitor elusive lizards: comparison of capture-recapture methods on giant day geckos (Gekkonidae, Phelsuma madagascariensis grandis) in the Masoala rainforest exhibit, Zurich Zoo",
    abstract = "Rapid and reliable estimation of population size is needed for the efficient monitoring of animal populations of conservation concern. Unfortunately, technical advances in this area have not been paralleled in uptake in conservation, which may be due to difficulties in implementation or the lack of general guidelines for application. Here we tested five different methods used to estimate population size [capture-mark-recapture (CMR), finite-mixture models, model averaging of finite-mixture models, accumulation curve methods (ACM), and the line transect method (LT)] using extensive capture{\^a}¿¿recapture data of the giant day gecko (Gekkonidae, Phelsuma madagascariensis grandis, Gray 1870) at the Masoala rainforest exhibit, Zurich Zoo. When the complete data were analyzed [30 sessions (and 27 sessions for the LT)], all methods except the LT produced similar estimates of population size. The simple ACM gave a small coefficient of variation (CV), but did not cover the most likely value of population size at moderate sampling effort. Nevertheless, the ACM was the only method that showed a reasonable convergence when subsets of data were used. CMR and Pledger models included the reference value in their confidence intervals (CI) after 25 and 30 sessions, respectively. Although model averaging did slightly improve the estimate, the CV was still high for the full dataset. Our method of using subsets of data to test the robustness of estimates is simple to apply and could be adopted more widely in such analyzes to evaluate sensitivity to method of evaluation. In conclusion, simple accumulation methods showed similar efficiency to more complex statistical models, and are likely to be sufficiently precise for most conservation monitoring purposes.",
    keywords = "Capture-mark-recapture, Finite-mixture models, Model averaging, Accumulation curve method, Line transect method, Method comparison.",
    author = "Thomas Wanger and Iris Motzke and Samuel Furrer and Barry Brook and Bernd Gruber",
    year = "2009",
    doi = "10.1007/s11284-008-0511-3",
    language = "English",
    volume = "24",
    pages = "345--353",
    journal = "Ecological Research",
    issn = "0912-3814",
    publisher = "Springer",

    }

    How to monitor elusive lizards: comparison of capture-recapture methods on giant day geckos (Gekkonidae, Phelsuma madagascariensis grandis) in the Masoala rainforest exhibit, Zurich Zoo. / Wanger, Thomas; Motzke, Iris; Furrer, Samuel; Brook, Barry; Gruber, Bernd.

    In: Ecological Research, Vol. 24, 2009, p. 345-353.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - How to monitor elusive lizards: comparison of capture-recapture methods on giant day geckos (Gekkonidae, Phelsuma madagascariensis grandis) in the Masoala rainforest exhibit, Zurich Zoo

    AU - Wanger, Thomas

    AU - Motzke, Iris

    AU - Furrer, Samuel

    AU - Brook, Barry

    AU - Gruber, Bernd

    PY - 2009

    Y1 - 2009

    N2 - Rapid and reliable estimation of population size is needed for the efficient monitoring of animal populations of conservation concern. Unfortunately, technical advances in this area have not been paralleled in uptake in conservation, which may be due to difficulties in implementation or the lack of general guidelines for application. Here we tested five different methods used to estimate population size [capture-mark-recapture (CMR), finite-mixture models, model averaging of finite-mixture models, accumulation curve methods (ACM), and the line transect method (LT)] using extensive captureâ¿¿recapture data of the giant day gecko (Gekkonidae, Phelsuma madagascariensis grandis, Gray 1870) at the Masoala rainforest exhibit, Zurich Zoo. When the complete data were analyzed [30 sessions (and 27 sessions for the LT)], all methods except the LT produced similar estimates of population size. The simple ACM gave a small coefficient of variation (CV), but did not cover the most likely value of population size at moderate sampling effort. Nevertheless, the ACM was the only method that showed a reasonable convergence when subsets of data were used. CMR and Pledger models included the reference value in their confidence intervals (CI) after 25 and 30 sessions, respectively. Although model averaging did slightly improve the estimate, the CV was still high for the full dataset. Our method of using subsets of data to test the robustness of estimates is simple to apply and could be adopted more widely in such analyzes to evaluate sensitivity to method of evaluation. In conclusion, simple accumulation methods showed similar efficiency to more complex statistical models, and are likely to be sufficiently precise for most conservation monitoring purposes.

    AB - Rapid and reliable estimation of population size is needed for the efficient monitoring of animal populations of conservation concern. Unfortunately, technical advances in this area have not been paralleled in uptake in conservation, which may be due to difficulties in implementation or the lack of general guidelines for application. Here we tested five different methods used to estimate population size [capture-mark-recapture (CMR), finite-mixture models, model averaging of finite-mixture models, accumulation curve methods (ACM), and the line transect method (LT)] using extensive captureâ¿¿recapture data of the giant day gecko (Gekkonidae, Phelsuma madagascariensis grandis, Gray 1870) at the Masoala rainforest exhibit, Zurich Zoo. When the complete data were analyzed [30 sessions (and 27 sessions for the LT)], all methods except the LT produced similar estimates of population size. The simple ACM gave a small coefficient of variation (CV), but did not cover the most likely value of population size at moderate sampling effort. Nevertheless, the ACM was the only method that showed a reasonable convergence when subsets of data were used. CMR and Pledger models included the reference value in their confidence intervals (CI) after 25 and 30 sessions, respectively. Although model averaging did slightly improve the estimate, the CV was still high for the full dataset. Our method of using subsets of data to test the robustness of estimates is simple to apply and could be adopted more widely in such analyzes to evaluate sensitivity to method of evaluation. In conclusion, simple accumulation methods showed similar efficiency to more complex statistical models, and are likely to be sufficiently precise for most conservation monitoring purposes.

    KW - Capture-mark-recapture

    KW - Finite-mixture models

    KW - Model averaging

    KW - Accumulation curve method

    KW - Line transect method

    KW - Method comparison.

    U2 - 10.1007/s11284-008-0511-3

    DO - 10.1007/s11284-008-0511-3

    M3 - Article

    VL - 24

    SP - 345

    EP - 353

    JO - Ecological Research

    JF - Ecological Research

    SN - 0912-3814

    ER -