A likelihood approach to estimating animal density from binary acoustic transects

Area of Science:

• Ecology
• Acoustic monitoring
• Statistical modeling

Background:

• Estimating animal density and abundance is crucial for ecological research and conservation.
• Passive acoustic monitoring provides valuable data but presents challenges in estimating detection parameters.
• Existing methods often require prior knowledge of detection probability and range, limiting their applicability.

Purpose of the Study:

• To develop an approximate maximum likelihood method for estimating animal density and abundance from binary passive acoustic transects.
• To simultaneously estimate unknown probability of detection and range of detection.
• To address the challenge of dependent data points in transect surveys.

Main Methods:

• Utilized an approximate maximum likelihood approach for data analysis.
• Assumed a homogeneous Poisson process in space for the data.
• Employed a second-order Markov approximation to the likelihood function.
• Exploited the dependence between successive data points in the transect survey.

Main Results:

• The proposed method demonstrated small bias under its derived assumptions.
• Performance was notably better when the probability of detection approached 1.
• The method's sensitivity to spatial trends in density and data clustering was identified.
• The approach was successfully illustrated using real acoustic data from whale surveys.

Conclusions:

• The developed method offers a viable approach for estimating animal density and abundance from passive acoustic transects.
• It effectively handles unknown detection probability and range, enhancing ecological survey capabilities.
• Researchers should be mindful of the method's sensitivity to spatial heterogeneity and clustering in data.