Generic variance expressions, precision, and sampling optimization for the sex-age-kill model of population reconstruction

John R. Skalski, Joshua J. Millspaugh

    Research output: Contribution to journalArticlepeer-review

    Abstract

    Many analytical techniques have been developed to estimate wildlife abundance from harvest data. The Sex-Age-Kill model of population reconstruction has a long history of use to monitor white-tailed deer (Odocoileus virginianus) numbers. The model is readily interpretable using data commonly collected by wildlife agencies. These features make implementation straightforward and abundance estimates understandable. However, the nonlinear nature of the estimation scheme complicates the variance structure of the model and how individual input parameters influence the precision of abundance estimates. We present generic variance expressions for adult male, adult female, juvenile, and total population abundance derived for the Sex-Age-Kill model. Using these variance expressions, we illustrate how the precision of the input parameters influences the overall precision of abundance estimates. To obtain the "rough management" benchmark (Robson and Regier 1964) for the precision of the abundance estimates, the individual input parameters must be estimated with coefficients of variation (CVs) < 0.15. An intermediate level of precision for "accurate management" would require CVs < 0.08; to achieve a high level of precision for "careful research" would constrain all input parameter CVs < 0.03. These results suggest that very precise field data are required for reliable population estimates. We also develop a cost function that identifies how funds should be spent to improve model estimates most economically. We suggest that researchers examine other heuristic population assessment techniques to develop rigorous methods for survey design and data analysis.

    Original languageEnglish
    Pages (from-to)1308-1316
    Number of pages9
    JournalJournal of Wildlife Management
    Volume66
    Issue number4
    DOIs
    StatePublished - Oct 2002

    Keywords

    • Cost function
    • Elk
    • Field sampling
    • Harvest
    • Odocoileus virginianus
    • Population analysis
    • Population assessment
    • Population reconstruction
    • Precision
    • Sampling
    • Sex-Age-Kill
    • Variance
    • White-tailed deer
    • Wisconsin method

    Fingerprint

    Dive into the research topics of 'Generic variance expressions, precision, and sampling optimization for the sex-age-kill model of population reconstruction'. Together they form a unique fingerprint.

    Cite this