A bayesian approach to parameter estimation for a crayfish (Procambarus spp.) bioaccumulation model.

Area of Science:

• Environmental toxicology
• Ecological risk assessment
• Computational toxicology

Background:

• Traditional bioaccumulation models use average parameter values, ignoring uncertainty and variability in predictions.
• Probabilistic methods, like Bayesian analysis, are recommended by the U.S. Environmental Protection Agency for quantifying uncertainty in ecological and human health risk estimates.

Purpose of the Study:

• To apply a Bayesian analysis to a crayfish bioaccumulation model, integrating uncertainty and variability of model parameters.
• To predict the distribution of chrysene concentration over time in crayfish from the LaBranche Wetlands.

Main Methods:

• A Bayesian analysis was performed using Markov chain Monte Carlo (MCMC) simulation.
• Experimental data on polycyclic aromatic hydrocarbon concentrations and lipid fractions in crayfish after a 5-day exposure were utilized.
• Posterior distributions of model parameters were derived from joint posterior distributions and experimental data.

Main Results:

• The Bayesian approach successfully accounted for combined uncertainty and variability in the crayfish bioaccumulation model parameters.
• Predicted ranges of chrysene concentration versus time were generated for crayfish at different study sites.
• This probabilistic approach offers a more comprehensive understanding of chemical uptake dynamics.

Conclusions:

• Bayesian analysis provides a robust framework for bioaccumulation modeling, improving the accuracy of ecological risk assessments.
• The study demonstrates the utility of probabilistic methods in ecological hazard evaluations.
• This approach enhances the prediction of chemical contaminant levels in aquatic organisms.