Statistical representativeness of ballast water sampling

Abstract

The International Maritime Organization's (IMO's) G2 guideline (Guidelines for ballast water sampling) clearly states that ballast water samples are required to be representative of the whole ballast water discharged. Enforcement and implementation of the ballast water convention largely depend on defendable case based on evidence that a ship entering territorial waters has not complied with regulations. The statistical representativeness of ballast water samples has been discussed in this study with additional emphasis on practicality of the sampling procedure. The mathematical formula for determining a sample size for a proportion from a finite population has been used to identify the minimum number of samples which could be considered as statistically representative of the ship's ballast water. Results clearly indicate that a large amount of ballast water, the 16 per cent of ballast water (8000 m(3)) for a ship discharging 50 000 m(3) of ballast water, must be sampled to achieve certain level of confidence which could be translated as a true representation of the ship's ballast water discharged and potentially used in any further legal actions by states or ship operators.