Application of Multiple Linear Regression Models in Deriving Water Quality Criteria for Lead: Comparison of Metal Bioavailability Estimation Approaches

Developing site-specific water quality criteria (WQC) that account for metal bioavailability can help accurately assess and control the ecological risks of lead (Pb) in water. However, balancing transparency and accuracy remains a challenge for conventional metal bioavailability estimation approaches based on water hardness and the biotic ligand model (BLM). Therefore, based on a recompiled toxicity data set of Pb, this study developed multiple linear regression (MLR) models incorporating key independent variables of pH, dissolved organic carbon (DOC), and hardness using the stepwise regression method. Results indicated that the BLM and pooled MLR models demonstrated similarities in toxicity prediction performance and residual distribution patterns for the same species. In terms of deriving WQC for Pb, the use of hardness-based WQC tended to be excessively stringent, particularly at low hardness levels. Generally, MLR- and BLM-based WQC were comparable across the distinct water quality conditions, despite variations in the responses of WQC derived from the two approaches to water quality parameters, particularly pH. The most notable disparities between WQC derived from the two approaches occurred under conditions characterized by extremely low DOC. In such instances, the MLR-based approach was recommended as an alternative to the BLM-based approach for deriving WQC for Pb.