Extensions of the Underhill-Zucchini Model for Avian Molt Data Allowing Random Duration

Abstract

Avian molting process is worth studying for its role in reflecting climate change or human activities. Two quantities, initiation molt time and duration time, are of great significance when describing avian molt. Existing models for the molting process use either a linear-based model, which treats both parameters as constants for all birds, or the Underhill-Zucchini model, which treats the initiation molt time as a random variable. In this thesis, we study extended models that allow both initiation time and duration to be random. The identifiability of the new model and its relationship with other models are then assessed. Using simulation studies, we compare the results from various models and evaluate conditions when some models are preferable. Finally, we show that we can get more accurate estimates from study designs with multiple observations per individual.