A Weight-of-Evidence Framework for Assessing Sediment (Or Other) Contamination: Improving Certainty in the Decision-Making Process

A basic framework is presented for the ecological weight-of-evidence (WOE) process for sediment assessment that clearly defines its essential elements and will improve the certainty of conclusions about whether or not impairment exists due to sediment contamination, and, if so, which stressors and biological species (or ecological responses) are of greatest concern. The essential “Certainty Elements” are addressed in a transparent best professional judgment (BPJ) process with multiple lines-of-evidence (LOE) ultimately quantitatively integrated (but not necessarily combined into a single value). The WOE Certainty Elements include: (1) Development of a conceptual model (showing linkages of critical receptors and ecosystem quality characteristics); (2) Explanation of linkages between measurement endpoint responses (direct and indirect with associated spatial/temporal dynamics) and conceptual model components; (3) Identification of possible natural and anthropogenic stressors with associated exposure dynamics; (4) Evaluation of appropriate and quantitatively based reference (background) comparison methods; (5) Consideration of advantages and limitations of quantification methods used to integrate LOE; (6) Consideration of advantages and limitations of each LOE used; (7) Evaluation of causality criteria used for each LOE during output verification and how they were implemented; and (8) Combining the LOE into a WOE matrix for interpretation, showing causality linkages in the conceptual model. The framework identifies several statistical approaches for integrating within LOE, the suitability of which depends on physical characteristics of the system and the scale/nature of impairment. The quantification approaches include: (1) Gradient (regression methods); (2) Paired reference/test (before/after control impact and ANOVA methods); (3) Multiple reference (ANOVA and multivariate methods); and 4) Gradient with reference (regression, ANOVA and multivariate methods). This WOE framework can be used for any environmental assessment and is most effective when incorporated into the initial and final study design stages (e.g., the Problem Formulation and Risk Characterization stages of a risk assessment) with reassessment throughout the project and decision-making process, rather than in a retrospective data analysis approach where key certainty elements cannot be adequately addressed.