1st Edition

Application of Uncertainty Analysis to Ecological Risks of Pesticides

Edited By William J. Warren-Hicks, Andy Hart Copyright 2010
    228 Pages 46 B/W Illustrations
    by CRC Press

    228 Pages 46 B/W Illustrations
    by CRC Press

    While current methods used in ecological risk assessments for pesticides are largely deterministic, probabilistic methods that aim to quantify variability and uncertainty in exposure and effects are attracting growing interest from industries and governments. Probabilistic methods offer more realistic and meaningful estimates of risk and hence, potentially, a better basis for decision-making. Application of Uncertainty Analysis to Ecological Risks of Pesticides examines the applicability of probabilistic methods for ecological risk assessment for pesticides and explores their appropriateness for general use.

    The book presents specific methods leading to probabilistic decisions concerning the registration and application of pesticides and includes case studies illustrating the application of statistical methods. The authors discuss Bayesian inference, first-order error analysis, first-order (non-hierarchical) Monte Carlo methods, second-order Bayesian and Monte Carlo methods, interval analysis, and probability bounds analysis. They then examine how these methods can be used in assessments for other environmental stressors and contaminants.

    There are many methods of analyzing variability and uncertainty and many ways of presenting the results. Inappropriate use of these methods leads to misleading results, and experts differ on what is appropriate. Disagreement about which methods are appropriate will result in wasted resources, conflict over findings, and reduced credibility with decision makers and the public. There is, therefore, a need to reach a consensus on how to choose and use appropriate methods, and to present this in the form of guidance for prospective users. Written in a clear and concise style, the book examines how to use probabilistic methods within a risk-based decision paradigm.

    Introduction and Objectives, A. Hart, D. Farrar, D. Urban, D. Fischer, T. La Point,
    K. Romijn, and S. Ferson
    Introduction
    Variability and Uncertainty
    Importance of Variability and Uncertainty in Risk Assessment
    Current Methods for Dealing with Variability and Uncertainty Are Inadequate
    Variability and Uncertainty Hinder the Regulatory Process
    Understanding Uncertainty and Variability Is Critical When Developing a Credible Risk Assessment
    Quantitative Analysis of Variability and Uncertainty Can Help
    When Is Quantitative Analysis of Variability and Uncertainty Required?
    What If the Bounds Are Very Wide?
    Need for Consensus on Appropriate Methods
    Workshop Objectives and Key Issues
    References

    Problem Formulation for Probabilistic Ecological Risk Assessments, A. Hart, S. Ferson, J. Shaw, G. W. Suter II, P. F. Chapman, P. L. de Fur, W. Heger, and P. D. Jones
    Introduction
    Main Steps in Problem Formulation
    Integration of Available Information for Probabilistic Assessments
    Definition of Assessment Endpoints for Probabilistic Assessments
    Definition of Assessment Scenarios
    Developing Conceptual Models for Probabilistic Assessments
    Analysis Plans for Probabilistic Assessment
    References

    Issues Underlying the Selection of Distributions, D. Farrar, T. Barry, P. Hendley, M. Crane, P. Mineau, M. H. Russell, and E. W. Odenkirchen
    Introduction
    Technical Background
    Some Practical Aspects of the Selection of Univariate Distributions
    Using Scanty and Fragmentary Data
    References

    Monte Carlo, Bayesian Monte Carlo, and First-Order Error Analysis, W. J. Warren-Hicks, S. Qian, J. Toll, D. L. Fischer, E. Fite, W. G. Landis, M. Hamer, and E. P. Smith
    Introduction
    Practical Aspects of a Monte Carlo Analysis
    Mathematical and Statistical Underpinnings of Monte Carlo Methods
    Bayesian Monte Carlo Analysis
    First-Order Error Analysis
    A Monte Carlo Case Study: Derivation of Chronic Risk Curves for Atrazine in Tennessee Ponds Using Monte Carlo Analysis
    Conclusions
    References

    The Bayesian Vantage for Dealing with Uncertainty, D. A. Evans, M. C. Newman, M. Lavine, J. S. Jaworska, J. Toll, B. Brooks, and T. C. M. Brock
    Introduction
    Conventional (Frequentist) Inference Methods
    Experiments Change the State of Knowledge
    Rules of Probability
    Bayes’s Theorem
    Examples Relevant to Uncertainty in Risk Assessment Quantifying Plausibility of a Cause–Effect Model
    Conclusion
    References

    Bounding Uncertainty Analyses, S. Ferson, D. R. J. Moore, P. Van den Brink, T. L. Estes, K. Gallagher, R. O’Connor, and F. Verdonck
    Introduction
    Robust Bayes
    Probability Bounds Analysis
    Numerical Example
    How to Use Bounding Results
    Seven Challenges in Risk Analyses
    What Bounding Cannot Do
    Example: Insectivorous Birds’ Exposure to Pesticide
    Conclusion
    Appendix
    References

    Uncertainty Analysis Using Classical and Bayesian Hierarchical Models, D. R. J. Moore, W. J. Warren-Hicks, S. Qian, A Fairbrother, T. Aldenberg, T. Barry, R. Luttik, and H.-T. Ratte
    Introduction
    Variability and Uncertainty
    Simple 2nd-Order Monte Carlo Analysis Case Study
    Bayesian Hierarchical Modeling
    References

    Interpreting and Communicating Risk and Uncertainty for Decision Making, J. L. Shaw, K. R. Tucker, K. Aden, J. M. Giddings, D. M. Keehner, and C. Kriz
    Introduction
    Participants in Risk Communication
    Communicating Uncertainty to Stakeholders and Participants
    Process for Communication
    Risk Assessor and Decision Maker Roles and Responsibilities
    Communication of Uncertainty for Regulatory Decision Making
    References

    How to Detect and Avoid Pitfalls, Traps, and Swindles, G. Joermann, T. W. La Point, L. A. Burns, J. P. Carbone, P. D. Delorme, S. Ferson, D. R. J. Moore, and T. P. Traas
    Introduction
    Meaningful Problem Formulation
    Suitability of Input Data
    Parameterization of the Distribution of Input Variables
    Correlations and Dependencies
    Model Uncertainties
    Software Tools and Computational Issues
    Presentation and Interpretation of Results
    Conclusions
    References

    Conclusions, A. Hart, T. Barry, D. L. Fischer, J. M. Giddings, P. Hendley, G. Joermann, R. Luttik, D. R. J. Moore, M. C. Newman, E. Odenkirchen, and J. L. Shaw
    Introduction
    Which Methods of Uncertainty Analysis Are Appropriate under What Circumstances?
    What Are the Implications of Probabilistic Methods for Problem Formulation?
    How Can Uncertainty Analysis Methods Be Used Efficiently and Effectively in Decision Making?
    When and How Should We Separate Variability and Uncertainty?
    How Can We Take Account for Uncertainty Concerning the Structure of the Risk Model for the Assessment?
    How Should We Select and Parameterize Input Distributions When Data Are Limited?
    How Should We Deal with Dependencies, Including Nonlinear Dependencies and Dependencies about Which Only Partial Information Is Available?
    How Can We Take Account of Uncertainty When Combining Different Types of Information in an Assessment (e.g., Quantitative Data and Expert Judgment, Laboratory Data, and Field Data)?
    How Can We Detect and Avoid Misleading Results?
    How Can We Communicate Methods and Outputs Effectively to Decision Makers and Stakeholders?
    What Are the Priorities for Further Development, Implementation, and Training?
    References
    Glossary

    Biography

    William J. Warren-Hicks, Andy Hart