3.1 All samples and analysis carried out to determine whether a ship is in
compliance with the BWM Convention should be performed under reliable and verified
QA/QC procedures (note that any method, approach or sampling procedure should be
rigorously validated and practicability should be assessed).
3.2 The first premise of any sampling and/or any analysis protocol is to
identify the purpose of the protocol, i.e. to prove whether the discharge of a ship is
meeting the D-1 standard or meeting the D-2 standard. There are many ways in which this
can be done; however, they are limited by:
-
.1 the requirements of the methodologies available for sampling the ballast water
discharge;
-
.2 the methods of analysis of samples being collected;
-
.3 the methods involved in statistically processing the results of these analyses;
-
.4 the specific operation of the ballast water management system (including when
the treatment is applied during the ballast cycle and the type of treatment used);
and
-
.5 the practicalities of sampling a very large volume of water and analysing it
for very low concentrations of organisms.
3.3 Successful sampling and analysis is also based on identifying the
viable biological population being sampled and its variability. If this population is
homogenous, it is much easier to sample than one that is known to be heterogeneous. In
the case of ballast water, the sample is drawn from a discharge with a population that
can vary significantly. Consequently, the samples collected for indicative or detailed
analysis should be representative samples.
3.4 Sampling a ballast water discharge is restricted even further when parts
of the ballast water may have already been discharged. Very few inferences can be made
on the quality of that ballast water already discharged based on sampling the remaining
discharge as it happens. The challenge is to determine the volume of interest and how to
sample it.
3.5 The qualitative difference between indicative analysis and detailed analysis often
relies on the level of statistical confidence, which, in detailed analysis may be
superior.
3.6 Indicative analysis (using operational or performance indicators) can
be undertaken at any time throughout the discharge. In cases where indicative analysis
identifies that a system is grossly exceeding the D-2 standard, it may be sufficient to
establish non-compliance, however, the practicalities, application and limitations of
the methodology being used for indicative analysis need to be understood fully.
3.7 Based on the discussion in paragraph 3.3, two different potential detailed sampling
approaches can therefore be considered:
-
.1 sampling the entire discharge from a vessel during a port visit. During this
approach:
-
.1 it will be impossible, by definition, for vessels to discharge prior to
sampling;
-
.2 large numbers of samples are likely to be required over a long period of
time;
-
.3 large sample volumes may be required over a long period of time; and
-
.4 sampling personnel would be required on the vessel over a significant
period of time; and
-
.2 collecting a representative sample of the ballast water being discharged during
some chosen period of time, e.g. one sample or a sequence of samples. During this
approach:
-
.1 the sampling can be developed to fit the situation on board the vessel;
and
-
.2 a representative sample of the discharge can be taken, and
that volume can be selected in many ways, providing the opportunity for
identifying and sampling specific volumes of the discharge if appropriate,
e.g. choosing a percentage of the discharge or sampling duration.
3.8 The D-2 standard expresses a low concentration of organisms to identify
in the analysis. The confidence in the result of any sampling and analysis depends on
the error inherent in the sampling method and on the error inherent in the method used
for analysing the sample. The cumulative error of both must be taken into account when
evaluating the result.
3.9 The tables in sections 4.1, 4.2 and 4.3 set out the range of methodologies and
approaches, currently identified for use to analyse ballast water discharges and how
they relate to the specific sampling protocols in section 4.4. These methodologies and
approaches are stand-alone techniques that need to be combined with specific sampling
protocols. These protocols should recognize the limitations of each methodology, its
inherent sampling requirements, and how it can fit into a comprehensive sampling
protocol for compliance testing.
3.10 Although some methodologies and approaches used in type approval testing may also
be applicable in compliance testing, the latter, especially indicative sampling, may
also require other approaches.