How to Recognize Your Pachyderm in a Hurry: Morphological Trait Diagnosticity for the Identification of Adult Asian Elephants

Abstract

The ability to track individual animals facilitates elephant conservation. Here, we test a measure of morphological trait diagnosticity to improve the identification of individual Asian elephants. Data were archival trait-list descriptions of 72 adult elephants, extracted from camera trap videos by trained coders. Twenty-four traits were coded for each individual. We calculated the entropy of each trait as a measure of diagnosticity, then performed Monte Carlo simulations to produce point and interval estimates of the number of trait checks needed to identify a randomly selected individual elephant. Highest entropy traits were depigmentation on the body, ear folds, ear tears, and brush type. Simulations confirmed that checking traits in order of entropy allowed faster and more robust identification as compared to control conditions. Results imply that trait rankings based on entropy might be used to train novice coders to identify elephants.