Double-observer line transect surveys with Markov-modulated Poisson process models for animal availability

Area of Science:

• Wildlife biology
• Ecological statistics
• Survey methodology

Background:

• Line transect surveys are crucial for wildlife population estimation.
• Undetected animals, even when available, introduce bias.
• Existing models may not fully capture animal availability dynamics.

Purpose of the Study:

• Develop maximum likelihood methods for line transect surveys with zero-distance undetected animals.
• Incorporate animal availability using a Markov-modulated Poisson process.
• Enable simultaneous estimation of detection and availability parameters.

Main Methods:

• Developed maximum likelihood methods for line transect surveys.
• Incorporated a Markov-modulated Poisson process for animal availability.
• Included a mark-recapture component from double-observer surveys.
• Modeled scenarios with multiple detections and self-estimated availability parameters.

Main Results:

• Simulations showed improved estimator performance.
• Simultaneous estimation of parameters is feasible with multiple detections and double-observer data.
• Multiple detections significantly enhance precision for availability model parameters.
• Double-observer data improved estimator precision in a minke whale survey.

Conclusions:

• The developed methods accurately estimate detection probability in line transect surveys.
• Recording multiple detections of individuals is recommended for improved precision.
• Double-observer surveys offer substantial benefits for estimating animal availability and detection probability.