A multinomial model for estimating the size of a whale population from incomplete census data

Area of Science:

• Marine biology
• Wildlife population estimation
• Arctic ecology

Background:

• Accurate population size estimation is crucial for managing and conserving whale stocks.
• The western Arctic bowhead whale (Balaena mysticetus) population faces unique challenges due to its remote habitat and migratory patterns.

Purpose of the Study:

• To adapt and apply the removal method for estimating the population size of the western Arctic bowhead whale stock.
• To develop a statistical model that accounts for variable whale detection probabilities and incomplete observational data.

Main Methods:

• The study adapts the removal method, where whales counted at the first census camp are 'removed' from the population considered at the second camp.
• A trinomial model is used, with population size as the number of trials and counts from each camp as cell totals.
• The model is extended to incorporate varying visibility conditions by summing independent trinomial distributions and derives confidence intervals for incomplete observations.

Main Results:

• The adapted removal method provides a robust framework for estimating bowhead whale population size.
• The model successfully incorporates factors like visibility and incomplete data, leading to more reliable estimates.
• Confidence intervals are derived to address challenges posed by incomplete observations at census camps.

Conclusions:

• The removal method, adjusted for visibility and data completeness, is a viable approach for estimating western Arctic bowhead whale populations.
• This methodology enhances the accuracy of population estimates, supporting informed conservation and management decisions.
• Further refinement of statistical models can improve the precision of wildlife population assessments in challenging environments.