Comparing the applicability of ecological risk indices of metals based on PCA-APCS-MLR receptor models for ports surface sediments.

This study collected surface sediments from seven ports in Taiwan and analyzed their characteristics along with 10 metals. Enrichment factor (EF), relative EF (REF), potential ecological risk index (PERI), and mean effect range median quotient (m-ERM-q) were used to evaluate the levels of metal contamination and ecological risks in sediments. Principal component analysis (PCA) and the absolute principal component score-multiple linear regression (APCS-MLR) model were applied to quantify the main factors affecting the variations in sediment metals. The different normalization techniques that vary between indexes significantly affect the estimates of risk levels for sediment metals. APCS-MLR model confirmed the significant difference among the sediment quality indices in the degree of anthropogenic pollution, ranging in the order of REF (normalized with reference site and Fe, 97.0 %), PERI (normalized with reference site, 85.5 %), EF (normalized with crust and Fe, 79.4 %), and m-ERM-q (not normalized, 56.6 %).