JCSET | Watershed Studies Institute Research Symposium
Academic Level at Time of Presentation
Senior
Major
Mathematics
List all Project Mentors & Advisor(s)
Dr. Christopher Mecklin; Dr. Howard Whiteman
Presentation Format
Oral Presentation
Abstract/Description
Capture-recapture models are essential tools for estimating population dynamics in ecological studies. A fundamental component of these models is the capture history matrix, which records individual detection over time and serves as the basis for estimating survival and capture probabilities. This presentation explores three statistical approaches to these estimations: the Cormack-Jolly-Seber (CJS) model, the Hidden Markov Model (HMM) for CJS, and the Bayesian CJS model. The CJS model provides a likelihood-based framework for estimation, and the HMM CJS incorporates latent states into the model to account for uncertainty in detection. The Bayesian CJS extends this same analysis by integrating prior knowledge and using Markov Chain Monte Carlo (MCMC) methods. Applying these models to the capture data of the Arizona tiger salamander, I will compare the assumptions, estimation techniques, and results of these methods while also discussing their strengths and limitations in ecological research.
Spring Scholars Week 2025
Watershed Studies Institute Research Symposium
Predicting Capture and Survival Probabilities of the Arizona Tiger Salamander: A Comparison of Capture-Recapture Models
Capture-recapture models are essential tools for estimating population dynamics in ecological studies. A fundamental component of these models is the capture history matrix, which records individual detection over time and serves as the basis for estimating survival and capture probabilities. This presentation explores three statistical approaches to these estimations: the Cormack-Jolly-Seber (CJS) model, the Hidden Markov Model (HMM) for CJS, and the Bayesian CJS model. The CJS model provides a likelihood-based framework for estimation, and the HMM CJS incorporates latent states into the model to account for uncertainty in detection. The Bayesian CJS extends this same analysis by integrating prior knowledge and using Markov Chain Monte Carlo (MCMC) methods. Applying these models to the capture data of the Arizona tiger salamander, I will compare the assumptions, estimation techniques, and results of these methods while also discussing their strengths and limitations in ecological research.
