A method for estimating abundance of mobile populations using telemetry and counts of unmarked animals. Issue 10 (22nd October 2015)

• Record Type: Journal Article
• Authors: Clement, Matthew J.; O'Keefe, Joy M.; Walters, Brianne
• Is Part Of: Ecosphere. Volume 6:Issue 10(2015)
• Page Start: 1
• Page End: 13
• Publication Date: 2015-10-22
• Subjects: abundance estimator -- coincidence rate -- exit counts -- fission-fusion -- Indiana bats -- Myotis sodalis -- social groups -- surveys; Ecology -- Periodicals; Ecology; Periodicals; 577.05
• DOI: 10.1890/ES15-00180.1
• Languages: English
• ISSNs: 2150-8925

Abstract

Abstract : While numerous methods exist for estimating abundance when detection is imperfect, these methods may not be appropriate due to logistical difficulties or unrealistic assumptions. In particular, if highly mobile taxa are frequently absent from survey locations, methods that estimate a probability of detection conditional on presence will generate biased abundance estimates. Here, we propose a new estimator for estimating abundance of mobile populations using telemetry and counts of unmarked animals. The estimator assumes that the target population conforms to a fission‐fusion grouping pattern, in which the population is divided into groups that frequently change in size and composition. If assumptions are met, it is not necessary to locate all groups in the population to estimate abundance. We derive an estimator, perform a simulation study, conduct a power analysis, and apply the method to field data. The simulation study confirmed that our estimator is asymptotically unbiased with low bias, narrow confidence intervals, and good coverage, given a modest survey effort. The power analysis provided initial guidance on survey effort. When applied to small data sets obtained by radio‐tracking Indiana bats, abundance estimates were reasonable, although imprecise. The proposed method has the potential to improve abundance estimates for mobile species that have a fission‐fusion social structure, such as Indiana bats, because it does not condition detection on presence at survey locations and because it avoids certain restrictive assumptions.