Hidden process models for animal population dynamics

Area of Science:

• Ecology
• Population Dynamics
• Statistical Modeling

Background:

• Animal population dynamics are influenced by complex demographic and environmental factors.
• Accurate population estimates are often challenging due to inherent process variation and measurement errors.
• Understanding these factors is critical for effective wildlife management and conservation.

Purpose of the Study:

• To introduce and demonstrate the utility of hidden process models for ecological studies.
• To simultaneously address process variation and observation/estimation error in population assessments.
• To provide a framework for estimating true population abundances and dynamics parameters.

Main Methods:

• Utilizing stochastic and deterministic subprocesses to model birth, survival, maturation, and movement.
• Modeling observations as functions of true abundance with probability distributions for error.
• Employing computer-intensive methods like sequential Monte Carlo or Markov Chain Monte Carlo for parameter estimation.

Main Results:

• Demonstrated the formulation and fitting of a hidden process model.
• Successfully estimated underlying true population abundances and population dynamics parameters.
• Provided a practical application for analyzing complex population data.

Conclusions:

• Hidden process models are a powerful and practical tool for ecological research.
• This approach enhances the accuracy of population estimates by accounting for key sources of uncertainty.
• The methodology is applicable to various species, including the Sacramento River winter-run chinook salmon.