Optimal Schemes to Detect the Presence of Insect Species

• Published in: Conservation biology Vol. 17; no. 2; pp. 476 - 487
• Main Authors: Zonneveld, C., Longcore, T., Mulder, C.
• Format: Journal Article
• Language: English
• Published: Boston, MA, USA Blackwell Science, Inc 01.04.2003; Blackwell Science; Blackwell
• Subjects: Animal and plant ecology; Animal, plant and microbial ecology; Biological and medical sciences; Butterflies; Conservation biology; Diplomatic protocol; Endangered species; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Insect biology; Methods and techniques (sampling, tagging, trapping, modelling...); Mortality; Population size; Species; Statistical variance; Synecology; Terrestrial ecosystems; Wildlife conservation; Endangered species; Faunal survey; Arthropoda; Insecta; Lepidoptera; Probability; Method; Invertebrata; Detection; Mathematical model; Transect method; Optimization
• ISSN: 0888-8892; 1523-1739
• DOI: 10.1046/j.1523-1739.2003.01516.x

Survey protocols to detect the presence of particular species should be organized to make optimal use of available resources. Current protocols mostly deal with this problem intuitively, but the problem can also be addressed with mathematical modeling. To this aim, we used a model that describes the abundance of a single, discrete generation of adult insects. Furthermore, we assumed that year-to-year variation in timing of the flight period was attributable to variation in peak emergence only. We determined the spacing of a given number of survey days on a transect that minimized the chance of missing a species when it was actually present. This spacing was fixed to a Julian date and applied to all years. We also calculated the probability of detecting the species, which depends on the number of survey days. For most parameter values, 5 survey days sufficed to detect with high probability (≥0.95) populations with an observable population size (total number of insects that can be observed in the transect over the entire flight period) exceeding 10. For observable population sizes under 10, achieving a detection probability of 0.95 may require many more than 5 survey days. As an example of the method, we constructed a survey-detection scheme for the Quino checkerspot butterfly (Euphydryas editha quino), based on the limited information available about its flight period. For an observable population size over 11, this scheme had a detection probability of >0.95. We also provide detection schemes for species whose death rate and variation in emergence time within and between years can at present be characterized only as small, moderate, or large. These survey schemes should maximize species detection and reduce uncertainty about absence.