John Wiley & Sons, Ltd.

Probabilistic approaches to accounting for data variability in the practical application of bioavailability in predicting aquatic risks from metals

0
The Biotic Ligand Model (BLM) theoretically enables the derivation of Environmental Quality Standards that are based on true bioavailable fractions of metals. Several physico‐chemical variables (especially pH, major cations, DOC and dissolved metal concentrations) must, however, be assigned to run the BLM, but they are highly variable in time and space in natural systems. This paper describes probabilistic approaches for integrating such variability during the derivation of Risk Indexes. To describe each variable using a Probability Density Function (PDF), several methods were combined to (i) treat censored data (i.e., data below the Limit of Detection); (ii) incorporate the uncertainty of the solid‐to‐liquid partitioning of metals; and (iii) detect outliers. From a probabilistic perspective, two alternative approaches that are based on log‐normal and Gamma distributions were tested to estimate the probability of the PEC (Predicted Environmental Concentration) exceeding the PNEC (Predicted Non Effect Concentration), i.e., pPECPNEC>1. The probabilistic approach was tested on four real‐case studies based on copper‐related data collected from stations on the Loire and Moselle rivers. The approach described in this paper is based on BLM tools that are freely available for end‐users (i.e., the Bio‐Met software) and on accessible statistical data treatments. This approach could be used by stakeholders who are involved in risk assessments of metals for improving site‐specific studies. Integr Environ Assess Manag © 2013 SETAC

Customer comments

No comments were found for Probabilistic approaches to accounting for data variability in the practical application of bioavailability in predicting aquatic risks from metals. Be the first to comment!