Establishing large mammal population trends from heterogeneous count data.

Area of Science:

• Wildlife Biology
• Conservation Science
• Statistical Ecology

Background:

• Wildlife population monitoring is crucial for conservation but challenged by heterogeneous time series data from various census methods (e.g., aerial, ground, expert estimates).
• Existing methods struggle to integrate data from different sampling strategies, frequencies, and precision levels, hindering accurate trend analysis.

Purpose of the Study:

• To present a novel three-step method for modeling wildlife population trends from heterogeneous time series data.
• To develop a flexible and principled approach for analyzing diverse historical census data, enhancing conservation and management strategies.

Main Methods:

• A heuristic for constructing credible intervals for all animal count data, including those lacking precision measures.
• Development of conversion factors to standardize aerial and ground counts, with estimates provided for broad animal classes.
• Implementation of a Bayesian model to estimate population growth rates and trends from reconciled counts, incorporating species-specific demographic potential.

Main Results:

• The method successfully generates credible intervals for various count types and provides standardized conversion factors for different species classes.
• The Bayesian model accurately estimates population trends and annual rates of increase, accounting for data precision and demographic limits.
• A bespoke R package, popbayes, is developed for accessible implementation of the entire methodology.

Conclusions:

• The presented three-step method offers a robust solution for analyzing heterogeneous wildlife population time series data.
• The approach facilitates the principled use of all available historical census data, leading to more consistent and reliable trend estimates.
• This flexible method is applicable to a wide range of animal species and can be adapted to test hypotheses about population dynamics.