Cormack-Jolly-Seber model with environmental covariates: a P-spline approach

Area of Science:

• Ecology
• Statistical Modeling
• Wildlife Biology

Background:

• Capture-recapture models, like the Cormack-Jolly-Seber (CJS) model, are crucial for estimating animal population dynamics.
• These models often incorporate time-varying covariates (e.g., temperature, rainfall) but assume linear relationships.
• Non-linear relationships between covariates and survival/capture probabilities can lead to biased estimates.

Purpose of the Study:

• To extend the Cormack-Jolly-Seber (CJS) model by incorporating non-linear covariate effects.
• To develop a semi-parametric frequentist approach for modeling capture and survival probabilities.
• To evaluate the performance of the proposed method through simulations and real-world data.

Main Methods:

• Semi-parametric modeling of capture and survival probabilities using P-splines within a frequentist framework.
• Extension of the Cormack-Jolly-Seber (CJS) model to accommodate non-linear covariate relationships.
• Conducting simulation studies to assess estimator performance.
• Comparison with existing semi-parametric Bayesian approaches using simulated and real datasets.

Main Results:

• The P-spline approach effectively models non-linear relationships between covariates and capture-recapture parameters.
• Simulation studies demonstrate the reliability and accuracy of the proposed frequentist estimators.
• The new method shows comparable or improved performance against established Bayesian techniques.

Conclusions:

• Semi-parametric P-spline modeling offers a flexible and robust frequentist alternative for analyzing capture-recapture data with complex covariate effects.
• This approach enhances the accuracy of population estimates by better capturing environmental influences on survival and capture probabilities.
• The developed methods provide valuable tools for ecological and wildlife research.