Regression models for animal abundance estimation
(Y Wang with P Yip)
The study of animal abundance is often accompanied by capture-type methods, which rely on the capture of subsets of members of the population.
Homogenous capture effort has long been known to cause substantial bias in estimation. In the studies the capture probabilities are assumed to depend on covariate information, such as individual characteristics, environmental conditions and past capture history.
Non-parametric, parametric, semi-parametric and multiplicative-additive models are developed to estimate the animal abundance.
The methods were applied to data from the Mai Po Bird Sanctuary, Hong Kong.