Multivariate Risk Analysis of Echotoxic Chemicals of Ballast Water Chemicals Based on PCA and DSS Using ECOTOX GISIS Data

Abstract

This study proposes a multivariate risk classification model for ballast water treatment chemicals by integrating global datasets—ECOTOX (U.S. EPA) and GISIS (IMO). Using Principal Component Analysis (PCA), we analyze 37 substances based on acute toxicity (LC50), chronic toxicity (NOEC), and bioaccumulation potential (BCF). The aim is to provide a practical, data-driven tool to support ecological compliance, early warnings, and regulatory prioritization in maritime chemical management. Results show that 43.24% of substances fall into the high-risk category, while only 8.11% are low risk. PCA effectively reduces dimensionality, explaining 73.63% of variance with just two components. High-risk chemicals such as Dibromoacetic acid and Dichloroacetonitrile exhibit low NOEC and high BCF values—indicating significant ecotoxic potential, often underregulated. Some commonly used oxidants also reveal hidden chronic toxicity, suggesting gaps in current risk frameworks post-BWM Convention. We construct a risk-scoring matrix and chemical heatmap to visualize ecotoxic profiles, enabling real-time risk ranking and decision support. Unlike previous studies that focus solely on toxicity thresholds or narrative reviews, this approach integrates empirical data with decision logic to aid Port State Control (PSC) and IMO policy design. The method is replicable and adaptable to other maritime pollutants, especially in the ASEAN context, enhancing smart port readiness and ecological safeguarding.