Generic water quality criteria (WQC) of a chemical are usually set based on results generated from toxicity tests which were conducted using standard laboratory water with well-controlled physiochemical properties. However, in natural aquatic environments, physiochemical characteristics, such as salinity, total suspended solid, total organic carbon and the co-existence of chemical contaminants, often vary spatially and temporally. These parameters can, in turn, alter the bioavailability of target chemicals and, thus, influence their toxicity to marine organisms. To account for site specificity, the US Environmental Protection Agency's water-effect ratio (WER = site water-LC50 / laboratory water-LC50) procedure can be applied to derive site-specific WQC. Most past studies, however, were conducted for freshwater systems. Here, for the first time, the WER of copper (Cu) was determined for three marine water control zones (WCZs) in Hong Kong: Victoria Harbour, Deep Bay and Southern WCZs. Samples of water were collected from three locations within each WCZ, while acute toxicities to the marine diatom Skeletonema costatum, intertidal copepod Tigriopus japonicus and larvae of marine medaka Oryzias melastigma were determined in site or laboratory (artificial seawater) waters. Results of this study showed that conservative final WER relative coefficients for Cu ranged from 0.57 to 0.73 for the three WCZs, and water from some locations caused >30% mortality in the fish larvae in the controls (without Cu addition). These results suggested that current generic WQC for Cu are likely under-protective for marine organisms in the three areas, and it should be tightened by multiplying it with site-specific WER to offer better protection to marine biodiversity and integrity of the ecosystem.