Abstract:Hierarchical modeling is frequently used to model ecological processes because of its ability to handle complex ecological phenomena by decomposing them into naturally explainable sub-models. Hierarchical Bayesian approaches have gained widespread use in health, social, and environmental sciences, including in the estimation of demographic parameters such as survival. In this study, we combine Bayesian hierarchical models with acoustic telemetry data to estimate survival probabilities for high-latitude populations of an anadromous salmonid, the Arctic char, while incorporating environmental and biological covariates to assess their impact on survival. The model we present here can also account for temporally varying detection probabilities due to changes to the acoustic receiver array design and seasonal variation in the detection probabilities related to environmental conditions (e.g., ice vs. no ice). As previously documented in this species, survival was high (>0.87) and we found that the covariates pertaining to sea ice coverage and Fulton's condition factor impacted the survival probabilities. Contrary to our expectations, high-condition fish had lower survival rates. Survival was also considerably lower during the summer (open-water) compared to winter (ice-covered) seasons. While the biological explanations and implications of these findings require further exploration, they nonetheless demonstrate the utility of this approach. Specifically, we present a hierarchical Bayesian model that can consider environmental and biological covariates while accounting for varying detection probabilities, a major concern of acoustic telemetry studies. The model can be easily adapted for other taxa with similar life histories where mark recapture data are available and can be extended to include additional environmental (e.g., salinity) and biological parameters (e.g., sex).

A Bayesian Hidden Semi-Markov Model with Covariate-Dependent State Duration Parameters for High-Frequency Environmental Data

A Bayesian Hidden Semi-Markov Model with Covariate-Dependent State Duration Parameters for High-Frequency Data from Wearable Devices

Bayesian Nonparametric Hidden Semi-Markov Models

Bayesian Approximations to Hidden Semi-Markov Models for Telemetric Monitoring of Physical Activity

Infinite hidden markov models for multiple multivariate time series with missing data

Bayesian non-asymptotic extreme value models for environmental data

Estimating survival of an anadromous salmonid using Bayesian hierarchical models applied to acoustic telemetry, biological, and environmental data

A Bayesian multilevel hidden Markov model with Poisson-lognormal emissions for intense longitudinal count data

Stochastic Precipitation Generation for the Chesapeake Bay Watershed using Hidden Markov Models with Variational Bayes Parameter Estimation

Bayesian spatio-temporal models for stream networks

Bayesian continuous-time hidden Markov models with covariate selection for intensive longitudinal data with measurement error.

Bayesian Non-Homogeneous Markov Models via Polya-Gamma Data Augmentation with Applications to Rainfall Modeling

Analyzing trends in precipitation patterns using Hidden Markov model stochastic weather generators

A Hidden Absorbing Semi-Markov Model for Informatively Censored Temporal Data: Learning and Inference

AdaptSPEC-X: Covariate Dependent Spectral Modeling of Multiple Nonstationary Time Series

Bayesian Methods for Modeling Cumulative Exposure to Extensive Environmental Health Hazards

Variational Bayesian Variable Selection for High-Dimensional Hidden Markov Models

An Advanced Hidden Markov Model for Hourly Rainfall Time Series

Uncovering ecological state dynamics with hidden Markov models

A Hidden Semi-Markov Model with Duration-Dependent State Transition Probabilities for Prognostics

Continuous Time Hidden Markov Model for Longitudinal Data