Heard but not seen: an acoustic survey of the African forest elephant population at Kakum Conservation Area, Ghana

This study, designed to survey forest elephants (Loxodonta africana cyclotis) at Kakum Conservation Area, Ghana, is the first to apply acoustic methods to elephant abundance estimation and to compare results with independent survey estimates. Nine acoustic sensors gathered sound continuously for 38 days. Low-frequency calling rates have been established as useful elephant abundance indices at a Namibian watering hole and a central African forest clearing. In this study, we estimated elephant population size by applying an abundance index model and detection function developed in central Africa to data from simultaneous sampling periods on Kakum sensors. The sensor array recorded an average of 1.81 calls per 20-min sampling period from an effective detection area averaging 10.27 km(2). The resulting estimate of 294 elephants (95% CI: 259-329) falls within confidence bounds of recent dung-based surveys. An extended acoustic model, estimating the frequency with which elephants are silent when present, yields an estimate of 350 elephants (95% CI: 315-384). Acoustic survey confidence intervals are at least half as wide as those from dung-based surveys. This study demonstrates that acoustic surveying is a valuable tool for estimating elephant abundance, as well as for detecting other vocal species and anthropogenic noises that may be associated with poaching.