A Bayesian Model to Estimate Abundance Based on Scarce Animal Vestige Data

TL;DR

This paper introduces a Bayesian modeling framework that accurately estimates animal abundance from scarce vestige count data, reducing costs and risks compared to direct methods, and is effective even with minimal data.

Contribution

The study presents a novel Bayesian model for abundance estimation using trace counts, validated through simulations and multiple case studies, especially effective with limited data.

Findings

Models produce accurate abundance estimates with scarce data.

Effective across diverse species and geographic regions.

Reduces costs and risks compared to direct counting methods.

Abstract

We propose a modelling framework which allows for the estimation of abundances from trace counts. This indirect method of estimating abundance is attractive due to the relative affordability with which it may be carried out, and the reduction in possible risk posed to animals and humans when compared to direct methods for estimating animal abundance. We assess these methods by performing simulations which allow us to examine the accuracy of model estimates. The models are then fitted to several case studies to obtain abundance estimates for collared peccaries in Brazil, kit foxes in Arizona, red foxes in Italy and sika deer in Scotland. Simulation results reveal that these models produce accurate estimates of abundance at a range of sample sizes. In particular, this modelling framework produces accurate estimates when data is very scarce. The use of vestige counts in estimating abundance allows for the monitoring of species which may otherwise go undetected due to their reclusive nature. Additionally, the efficacy of these models when data is collected at very few transects will allow for the use of small-scale data collection programmes which may be carried out at reduced cost, when compared to larger-scale data collection.