Abundance estimation of diving animals by the double-platform line transect method

Area of Science:

• Ecology
• Wildlife Biology
• Statistical Ecology

Background:

• Conventional line transect theory often assumes perfect detection of animals on the transect line.
• This assumption can lead to biased abundance estimates in real-world ecological surveys.
• Accounting for imperfect detection and observer variability is crucial for accurate wildlife population assessments.

Purpose of the Study:

• To introduce an extended and generalized hazard probability model for line transect surveys.
• To relax the assumption of perfect detection, thereby improving abundance estimation.
• To incorporate random and systematic heterogeneity in detection probability.

Main Methods:

• Development of a generalized hazard probability model for line transect sampling.
• Utilizes a survey design with independent observers and distinguishes simultaneous/delayed duplicates.
• Incorporates random heterogeneity (e.g., surfacing behavior) and systematic heterogeneity (covariates).
• Employs the Horvitz-Thompson estimator for abundance estimation in unequal probability sampling.

Main Results:

• The proposed model effectively estimates animal abundance without assuming perfect detection.
• It successfully accounts for various sources of heterogeneity in detection probability.
• Simulation studies indicate the model performs well.
• The method was successfully applied to Antarctic minke whale data.

Conclusions:

• The generalized hazard probability model offers a more robust approach to line transect surveys.
• It provides improved abundance estimates by addressing imperfect detection and heterogeneity.
• The model's flexibility allows for extensions to various observer independence scenarios.
• This method enhances the accuracy of ecological population assessments.