Ballast Water Sampling and Sample Analysis for Compliance Control - Global Maritime Transport and Ballast Water Management: Issues and Solutions - page 187

Consequently the selection of inappropriate sampling approaches will infl uence the
compliance control result. In consequence, the concentration of viable organisms in the
ballast water discharge may be underestimated, so that ballast water managed with an
ineffi cient BWMS could be recognised as compliant. On the other hand, concentra-
tions of viable organisms may also be overestimated, and ballast water complying with
the D-2 standard may fail in compliance tests.
Sequential ballast water sampling trials documented different organism numbers
in each sequence of one test of the identical ballast water tank which indicates the
patchy organism distribution inside the tank. This was observed during all sampling
events conducted by Gollasch and David ( 2009 , 2010a , b , 2013 ) and for both organ-
ism groups studied. It was therefore concluded that sampling during ballast water
discharge is biased by tank patchiness of organisms.
For organisms greater than or equal to 50 m in minimum dimension our previ-

ously undertaken studies have shown that the samples taken over the entire dis-
charge time of a tank contained much lower concentrations of viable organism
compared to the organism count in the sequences. It was therefore concluded that
sequential sampling may deliver more representative results. Comparative studies
have further shown that sequential samples of approximately 10 min duration are
suitable for in-line D-2 standard compliance tests. In most of our tests the highest
count of viable zooplankton organisms was found in the last sequence so that sam-
pling at this time may “oversample” the real organism concentration. It was also
observed that sequential samples taken in the very beginning and very end during a
ballast tank is emptied are unlikely to provide representative results of the viable
organism concentrations because in these samples the organism count showed very
high variations. This could result in an under- or oversampling the organism con-
centration. Even when avoiding these time windows for sampling the concentration
of viable organisms still seems to be patchy and we therefore recommend to take at
least two sequential samples but excluding the very beginning and the very end of
the pumping times when a ballast tank is emptied. The mean value of the viable
organism concentration in these two sequential samples may be taken to assume the
real organism concentration.
In the group of organisms less than 50

ʼ

ʼ m in minimum dimension and greater

than or equal to 10 m in minimum dimension counts of viable individuals in

sequences when compared to the samples taken over the entire time of a tank dis-
charge showed lower viable organism concentrations in the sequential samples,
which is in contrast to the larger organism results. Comparisons of the smaller
organism concentration between the different sequences of all tests showed that no
clear trend can be identifi ed during which time window a more representative sam-
ple may be taken. It is therefore recommended to take at least two sequential sam-
ples during the discharge of a ballast water tank but avoiding sampling times during
the very beginning and end of the discharge of a tank or tanks. The mean organism
count in these two or more sequential samples may be seen as the real viable organ-
ism concentration in the ballast water.
Suggested methods and equipment are outlined in Table 5 and for a detailed
description of the sampling equipment and sampling arrangements see section
“ Sampling Equipment and Sampling Point Arrangements ”.

ʼ