Technical ReportPDF Available

Literature Review for the Indicative Ballast Water Analysis Methods.

Authors:

Abstract and Figures

The International Convention for the Control and Management of Ships' Ballast Water and Sediments was adopted by the International Maritime Organization (IMO) in 2004. The Convention will enter into force internationally in September 2017. The general obligations of the Ballast Water Management Convention include control measures that the Parties to the Convention are required to take to ensure that the ships entering their ports are in compliance with this Convention. In Finland, Trafi is the authority responsible for port state control inspections of ships. The inspection is primarily conducted as a documentary check; however, the authority may always carry out ballast water sampling to verify that the ship is in compliance with the Convention. The sampling consists of an indicative analysis and a detailed analysis. Indicative analysis refers to indicative sampling of the ballast water pumped out of a vessel. The results indicate whether the ship meets the performance standard laid down in the Ballast Water Management Convention. If the ship fails to meet the standard, a detailed analysis must be performed in a laboratory. Based on the laboratory results, it is decided whether further measures will be taken. The purpose of the study commissioned by Trafi was to find an indicative analysis method for the use of Trafi's port state control inspectors. The Finnish Environment Institute (SYKE) conducted the study for Trafi. Based on the study, three different methods were found to be best suited for the conditions in Finland and the Baltic Sea. The recommended methods are PAM (Pulse amplitude- modulation fluorometry), ATP (Adenosine triphosphate) and FRR (Fast repetition rate fluorometry). The most important assessment criteria were the reliability and user-friendliness of the method method, the time required for obtaining the results as well as the procurement and operating costs of the method.
Content may be subject to copyright.
Literature Review for the Indicative
Ballast Water Analysis Methods
Okko Outinen & Maiju Lehtiniemi
Trafi Publications
Trafin julkaisuja
Trafis publikationer
10/2017
Trafi Publications 10-2017
Title of publication
Literature Review for the Indicative Ballast Water Analysis Methods
Author(s)
Okko Outinen & Maiju Lehtiniemi, Finnish Environment Institute
Commissioned by, date
Finnish Transport Safety Agency, 11 October 2016
Publication series and number
Trafi Research Reports 10/2017
ISSN (online) 2342-0294
ISBN (online) 978-952-311-202-5
Keywords
Ballast Water Management Convention, Indicative analysis, Sampling
Contact person
Ville-Veikko Intovuori Language of the report
English
Abstract
The International Convention for the Control and Management of Ships' Ballast Water
and Sediments was adopted by the International Maritime Organization (IMO) in 2004.
The Convention will enter into force internationally in September 2017.
The general obligations of the Ballast Water Management Convention include control
measures that the Parties to the Convention are required to take to ensure that the
ships entering their ports are in compliance with this Convention. In Finland, Trafi is
the authority responsible for port state control inspections of ships. The inspection is
primarily conducted as a documentary check; however, the authority may always carry
out ballast water sampling to verify that the ship is in compliance with the Convention.
The sampling consists of an indicative analysis and a detailed analysis. Indicative anal-
ysis refers to indicative sampling of the ballast water pumped out of a vessel. The re-
sults indicate whether the ship meets the performance standard laid down in the Bal-
last Water Management Convention. If the ship fails to meet the standard, a detailed
analysis must be performed in a laboratory. Based on the laboratory results, it is de-
cided whether further measures will be taken. The purpose of the study commissioned
by Trafi was to find an indicative analysis method for the use of Trafi's port state con-
trol inspectors. The Finnish Environment Institute (SYKE) conducted the study for
Trafi.
Based on the study, three different methods were found to be best suited for the con-
ditions in Finland and the Baltic Sea. The recommended methods are PAM (Pulse am-
plitude-modulation fluorometry), ATP (Adenosine triphosphate) and FRR (Fast repeti-
tion rate fluorometry). The most important assessment criteria were the reliability and
user-friendliness of the method, the time required for obtaining the results as well as
the procurement and operating costs of the method.
Date
21 April 2017
Trafi Publications 10-2017
Julkaisun nimi
Literature Review for the Indicative Ballast Water Analysis Methods
Tekijät
Okko Outinen & Maiju Lehtiniemi, Suomen ympäristökeskus
Toimeksiantaja ja asettamispäivämäärä
Liikenteen turvallisuusvirasto, 11.10.2016
Julkaisusarjan nimi ja numero
Trafin tutkimuksia 10/2017
ISSN (verkkojulkaisu) 2342-0294
ISBN (verkkojulkaisu) 978-952-311-202-5
Asiasanat
Painolastivesiyleissopimus, indikatiivinen analyysi, näytteenotto
Yhteyshenkilö
Ville-Veikko Intovuori Raportin kieli
Englanti
Tiivistelmä
Alusten painolastivesien ja sedimenttien valvontaa ja käsittelyä koskeva kansainväli-
nen yleissopimus hyväksyttiin Kansainvälisessä merenkulkujärjestössä (IMO) vuonna
2004 ja se on tulossa kansainvälisesti voimaan syyskuussa 2017.
Painolastivesiyleissopimuksen keskeisiin velvoitteisiin kuuluu alusten vaatimustenmu-
kaisuuden valvonta, jota jäsenvaltioiden tulee tehdä aluksille niiden vieraillessa jäsen-
valtion satamissa. Suomessa vastuuviranomainen alusten satamavaltiotarkastuksissa
on Trafi. Tarkastus tehdään ensisijaisesti asiakirjatarkastuksena, mutta viranomaisella
on aina mahdollisuus suorittaa painolastivesinäytteenotto vaatimustenmukaisuuden to-
dentamiseksi.
Näytteenotto koostuu indikatiivisesta analyysista ja yksityiskohtaisesta analyysista. In-
dikatiivisella analyysillä tarkoitetaan suuntaa antavaa näytteenottoa aluksen ulospum-
pattavasta painolastivedestä. Tuloksista nähdään täyttääkö alus painolastivesiyleisso-
pimuksen mukaisen suorituskykystandardin. Mikäli näin ei ole, tulee suorittaa yksityis-
kohtainen analyysi, joka tehdään laboratoriossa. Laboratoriotulosten perusteella pää-
tetään ryhdytäänkö mahdollisiin jatkotoimenpiteisiin. Trafin teettämän selvityksen tar-
koituksena oli löytää indikatiivinen analyysimenetelmä Trafin satamavaltiotarkastajien
käyttöön. Selvityksen Trafille teki Suomen ympäristökeskus (SYKE).
Selvityksen perusteella löytyi kolme erilaista menetelmää, jotka parhaiten soveltuvat
Suomen ja Itämeren olosuhteisiin. Suositellut menetelmät olivat PAM- (Pulse ampli-
tude-modulation fluorometry), ATP- (Adenosine triphosphate) ja FRR (Fast repetition
rate fluorometry) -menetelmät. Tärkeimmät arviointikriteerit olivat menetelmän luotet-
tavuus, helppokäyttöisyys, ajallinen kesto tulosten saamiseksi sekä menetelmän han-
kinta- ja käyttökustannukset.
Julkaisun päivämäärä
21.4.2017
Trafi Publications 10-2017
Publikation
Literature Review for the Indicative Ballast Water Analysis Methods
Författare
Okko Outinen & Maiju Lehtiniemi, Finlands miljöcentral
Tillsatt av och datum
Trafiksäkerhetsverket
Publikationsseriens namn och nummer
Trafis undersökningsrapporter
10/2017
ISSN (webbpublikation) 2342-0294
ISBN (webbpublikation) 978-952-311-202-5
Ämnesord
Barlastvattenkonvention, indikativ analys, provtagning
Kontaktperson
Ville-Veikko Intovuori Raportens språk
finska
Sammandrag
Internationella sjöfartsorganisationen (IMO) antog 2004 den internationella konvent-
ionen om kontroll och hantering av fartygs barlastvatten och sediment. Konventionen
kommer att träda i kraft internationellt i september 2017.
Till de viktigaste skyldigheterna enligt barlastvattenkonventionen hör kontroll av att
fartygen uppfyller kraven, så medlemsstaterna ska kontrollera de fartyg som besöker
deras hamnar. I Finland är Trafi ansvarig myndighet för dessa hamnstatskontroller.
Kontrollen görs i första hand som en granskning av dokument, men myndigheten har
alltid möjlighet att ta prov av barlastvattnet för att konstatera att kraven uppfylls.
Provtagningen består av en indikativ analys och en detaljerad analys. Med indikativ
analys avses en riktgivande provtagning av det barlastvatten som pumpas ut ur farty-
get. Resultatet visar huruvida fartyget uppfyller de prestandanormer som föreskrivs i
barlastvattenkonventionen. Om så inte är fallet, ska en detaljerad analys utföras i la-
boratorium. Laboratorieresultaten avgör huruvida man vidtar eventuella vidare åtgär-
der. Syftet med utredningen som Trafi lät göra var att hitta en indikativ analysmetod
för Trafis hamnstatsinspektörer. Utredningen utfördes för Trafis räkning av Finlands
miljöcentral.
Utredningen fann tre olika metoder som lämpar sig bä
st för förhållandena i Finland och
Östersjön. De rekommenderade metoderna är PAM (Pulse amplitude-modulation
fluorometry), ATP (Adenosine triphosphate) och FRR (Fast repetition rate fluorometry).
De viktigaste bedömningskriterierna var att metoden är tillförlitlig, lätt att använda,
hur lång tid det tar innan resultaten är färdiga och vilka upphandlings- och driftskost-
naderna är.
Publikationsdatum
21.4.2017
Trafi Publications 10-2017
FOREWORD
The Finnish Transport Safety Agency (Trafi) has commissioned a study on the indicative
ballast water analysis methods. The results of the study have been compiled in this report.
The study was conducted as a literature review and was based on previous international
studies on indicative analysis methods, which were specified with expert interviews. The
aim of the study was to find the method best suited for the conditions in Finland and the
Baltic Sea. This is the first study of its kind conducted in Finland.
The study was carried out by Assisting Researcher Okko Outinen and Senior Researcher
Maiju Lehtiniemi of the Finnish Environment Institute (SYKE). The steering group of the
study included Special Adviser Ville-Veikko Intovuori, Chief Adviser Anita Mäkinen and
Head of Unit Mirja Ikonen of Trafi, and Okko Outinen and Maiju Lehtiniemi of SYKE.
Helsinki, 10 April 2017
Ville-Veikko Intovuori
Special Adviser
Finnish Transport Safety Agency (Trafi)
Trafi Publications 10-2017
ALKUSANAT
Liikenteen turvallisuusvirasto Trafi teetti tutkimuksen alusten painolastivesien indikatiivi-
sista analyysimenetelmistä, jonka tulokset on koottu tähän raporttiin. Tutkimus tehtiin kir-
jallisuuskatsauksena, perustuen aikaisempiin indikatiivisista analyysimenetelmistä tehtyihin
kansainvälisiin tutkimuksiin ja joita tarkennettiin asiantuntijahaastatteluilla. Tutkimuksen
tarkoitus oli löytää Suomen ja Itämeren olosuhteisiin parhaiten soveltuva menetelmä. Ai-
kaisemmin vastaavaa tutkimusta Suomessa ei ollut tehty.
Tutkimuksen tekivät Trafille Suomen ympäristökeskuksesta (SYKE) apulaistutkija Okko
Outinen ja erikoistutkija Maiju Lehtiniemi. Tutkimuksen ohjausryhmään osallistuivat Tra-
fista erityisasiantuntija Ville-Veikko Intovuori, johtava asiantuntija Anita Mäkinen ja yksi-
könpäällikkö Mirja Ikonen sekä SYKE:sta Okko Outinen ja Maiju Lehtiniemi.
Helsingissä, 10. huhtikuuta 2017
Ville-Veikko Intovuori
erityisasiantuntija
Liikenteen turvallisuusvirasto (Trafi)
Trafi Publications 10-2017
FÖRORD
Trafiksärkerhetsverket (Trafi) har låtit göra en undersökning om indikativa analysmetoder
av fartygs barlastvatten. Resultaten av undersökningen har sammanställts i denna rapport.
Undersökningen gjordes i form av en litteraturgenomgång som grundade sig på tidigare in-
ternationella undersökningar av indikativa analysmetoder och som kompletterades med ex-
pertintervjuer. Syftet med undersökningen var att hitta den metod som är mest lämplig med
tanke på förhållandena i Finland och Östersjön. Ingen motsvarande undersökning hade tidi-
gare gjorts i Finland.
Undersökningen utfördes för Trafis räkning av forskarassistent Okko Outinen och special-
forskare Maiju Lehtiniemi vid Finlands miljöcentral. I undersökningens styrgrupp deltog
specialsakkunnig Ville-Veikko Intovuori, ledande sakkunnig Anita Mäkinen och enhetschef
Mirja Ikonen från Trafi samt Okko Outinen och Maiju Lehtiniemi från Finlands miljöcen-
tral.
Helsingfors den 10 april 2017
Ville-Veikko Intovuori
specialsakkunnig
Trafiksäkerhetsverket (Trafi)
Trafi Publications 10-2017
Index
1. Introduction ................................................................................... 1
1.1 Testing compliance for Regulations D1 and D2.............................. 2
1.2 In-tank and in-line sampling for indicative analysis ....................... 4
1.3 Recommended principles for the D2 standard compliance testing .... 5
2. Materials and methods ................................................................... 7
3. Results............................................................................................ 8
3.1 Adenosine triphosphate (ATP) method ......................................... 8
3.1.1 Sampling approach ......................................................... 9
3.1.2 Feasibility .................................................................... 10
3.2 Fluorescein diacetate (FDA) method .......................................... 11
3.2.1 Sampling approach ....................................................... 11
3.2.2 Feasibility .................................................................... 13
3.3 Pulse amplitude-modulation (PAM) method ................................ 14
3.3.1 Sampling approach ....................................................... 15
3.3.2 Feasibility .................................................................... 16
3.4 Microscopy methods ................................................................ 17
3.4.1 Sampling approach and feasibility ................................... 17
3.5 Flow cytometry method ........................................................... 18
3.5.1 Sampling approach ....................................................... 18
3.5.2 Feasibility .................................................................... 20
3.6 Bacterial identification methods ................................................ 21
3.6.1 Detection of bacterial enzymes ....................................... 21
3.6.2 Real-time Polymerase chain reaction (PCR) ...................... 22
3.6.3 Colorimetric bacteria detection ....................................... 22
3.6.4 Other methods for detection of bacteria ........................... 23
3.7 Other potential indicative methods ............................................ 24
3.7.1 Microfluidic lab-on-chip biosensor .................................... 24
3.7.2 Fast repetition rate (FRR) fluorometry ............................. 26
3.7.3 Serial dilution culture-most probable number (SDC-MPN) method
.................................................................................. 27
3.7.4 Laser transmission spectroscopy (LTS) ............................ 28
3.7.5. Analysis techniques for larger zooplankton ....................... 28
3.8 Results overview ..................................................................... 29
4. Discussion .................................................................................... 31
4.1 General issues and uncertainties regarding the sampling methods 31
4.1.1 Minimally reliable methods ............................................. 32
4.1.2 Moderately reliable methods ........................................... 32
4.1.3 Highly reliable methods ................................................. 35
4.2 Accuracy ................................................................................ 35
4.3 Manufacturers ........................................................................ 37
4.4 Study limi tations ..................................................................... 39
5. Conclusions and recommendations ............................................... 39
6. Acknowledgements ....................................................................... 40
7. Bibliobraphy ................................................................................. 41
Appendices ........................................................................................ 49
Appendix A: SGS ATP sampling approach ............................................ 49
Appendix B: FDA Pulse counting device ............................................... 52
Appendix C: Indicative methods described in the present study .............. 53
Appendix D: The current target species list for the HELCOM area ............ 54
Trafi Publications 10-2017
1
1. Introduction
International Maritime Organization (IMO) adopted the International Convention for
the Control and Management of Ships Ballast Water and Sediments (BWM Conven-
tion) in 2004, to prevent the introductions of non-indigenous species (NIS) with two
separate BWM requirements for ships, ballast water exchange standard (Regulation
D1) and ballast water performance standard (Regulation D2) (Table 1) (IMO, 2009,
David et al., 2013). The transfer of NIS through ship’s ballast water and sediment
discharges is widely acknowledged as one of the most recent waterborne threats to
natural environment, human health and economy worldwide (Cordell et al., 2009,
David and Gollasch, 2015). Regulation D1 requires ships to exchange a minimum of
95% of the ballast water volume at the open sea. Regulation D2 in turn, requires
ships to conduct ballast water treatment in order to decrease the amount of dis-
charged viable organisms below the agreed limits (IMO, 2009, Albert et al., 2013).
The BWM Convention will enter into force 12 months after the date on which at
least 30 States, representing at least 35 % of the world’s merchant shipping tonnage
have ratified the Convention (IMO, 2009, Albert et al., 2013). Finland ratified the
convention on the 8th of September 2016, sufficiently pushing the tonnage percent-
age over the required limit. Therefore, as of 8th of September, 2016, 52 States have
ratified the BWM Convention representing 35.14 % of the world’s merchant ship-
ping tonnage and the BWM Convention will enter into force the 8th of September
2017 (IMO, 2016).
Table 1. Requirements for ballast water exchange and performance standards (IMO, 2009).
Regulation D1
Regulation D2
Sampling efficiency of at least 95% vol-
umetric exchange of ballast water. The
sampling standard is met, when the vol-
ume of each ballast water tank is
pumped through at least 3 times.
Ships conducting ballast water treatment shall
discharge:
1. less than 10 viable organisms per cubic me-
tre 50 micrometres (m) in minimum di-
mension,
2. Less than 10 viable organisms per millilitre
(ml) < 50 m in minimum dimension and 10
m in minimum dimension and
3. Discharge of the indicator microbes shall not
exceed:
a) Toxicogenic Vibrio cholerae with less than 1
colony forming unit (cfu) per 100 (ml) or less than
1 cfu per 1 gram of zooplankton samples, b)
Escherichia coli less than 250 cfu per 100 ml and
c) Intestinal Enterococci less than 100 cfu per
100 ml.
IMO has also provided general recommendations on methods and approaches for
compliance monitoring in terms of Regulations D1 and D2 of the BWM Conven-
tion. These recommendations are described in detail in the Guidance on ballast water
sampling and analysis for trial use in accordance with the BWM Convention and
Guidelines (G2, Resolution MEPC.173(58)) (IMO, 2015). Indicative ballast water
Trafi Publications 10-2017
2
sampling analyses have been recommended as preliminary tests for the determina-
tion of potential immediate mitigation measures on whether the ship is compliant or
non-compliant to discharge ballast water (MEPC, 2008). Successful ballast water
sampling for detailed compliance analysis includes the identification of viable or-
ganisms and their populations (IMO, 2015).
The purpose of this literature review was to provide a wide perspective on present
indicative analysis methods with special attention drawn to environmental condi-
tions and water characteristics in the coastal waters of the Baltic Sea. The purpose of
the study was supported with the following objectives;
To present all the existing indicative analysis methods for compliance with Regula-
tion D2 of the BWM Convention,
To evaluate the efficiency of each method in terms of lower water salinity and tem-
perature and relatively high water turbidity,
To determine and describe the principle of sampling, reliability, applicability, re-
quired skills, sampling time and required amount of ballast water, the price and man-
ufacturers and previous experiences with the testing for each analysis method and
To compare the advantages and disadvantages of the methods and provide recom-
mendations for the most suitable methods available for testing in practice.
1.1 Testing compliance for Regulations D1 and D2
Compliance for the ballast water exchange standard can be simply measured by de-
tecting the salinity of the ballast water (Gollasch and David, 2015). If the ship has
conducted ballast water exchange at the open sea, the salinity of the ballast water
should always be over 30. Water sample for salinity can be collected from the man-
hole, air vent, sounding pipe or discharge line and only 50 ml of water is required.
General methodological approaches for testing according to Regulation D1 compli-
ance are shown in Table 2. Pereira et al. (2016) also suggested that turbidity can be a
good indicator for the conduction of ballast water exchange. As ships trading in the
Baltic Sea do not have the option for ballast water exchange at open sea, the main
focus of this literature review is on testing for Regulation D2 compliance.
Table 2. General methodological approaches for testing Regulation D1 compliance (IMO, 2015).
Indicator
Sampling
ap-
proach
Notifications
Salinity Conductivity meter External elements can affect recorded salinity
Salinity Refractometer Varying temperature can alter the readings
Types of organ-
isms (Oce-
anic/coastal)
Visual identification Usually expensive, time-consuming and re-
quires extensive skills. Can lead to false re-
sults (encysted organisms from previous bal-
lasting operations hatch)
Turbidity Turbidity sensors Understanding of turbidity characteristics re-
quired
Dissolved or-
ganic and inor-
ganic constitu-
ents (nutrients,
Nutrient sensors Understanding of inorganic or organic constitu-
ent characteristics required
Trafi Publications 10-2017
3
metals, organic
matter)
Compliance for the ballast water performance standard can be detected by indicative
or detailed analysis (IMO, 2015). Indicative analysis refers to a relatively fast indi-
rect or direct measurement of a representative sample from the ballast water tank.
The main purpose of indicative sampling is to provide a quick estimation of the as-
sociated organism concentration and if the estimation indicates that the ship is non-
compliant, detailed analysis samples can be collected for further evaluation of the
compliance (David and Gollasch, 2015). Indirect measures can include physical,
chemical or biological parameters that require understanding of the method and indi-
cate changes in these parameters after potential ballast water treatment, whereas di-
rect measures are readily comparable to D2 standard (number of viable organ-
isms/volume) (IMO, 2015). A detailed analysis in turn can be defined as a direct
measurement of viable organism concentrations in the ballast water. Detailed analy-
sis is generally more complex than indicative analysis and tends to provide more
precise indications of the quality and quantity of the organism concentrations. The
main characteristics of these two analysis methods are represented in Table 3.
Table 3. The primary characteristics of indicative and detailed analysis methods (IMO, 2015).
Indicative analysis
Detailed analysis
Purpose
To deliver a quick, broad
estimation of the concen-
tration of viable organ-
isms
To deliver a more accurate, di-
rect measurement of the con-
centration of viable organisms
Sampling
volume
Varies and depends on the
method Varies and depends on the
method
Representative
sampling
Yes, represents organism
concentrations of volume of
interest
Yes, represents organism con-
centrations of volume of inter-
est
Analysis parameters
Can be operational (chemical,
physical), and/or indicate per-
formance (biological)
Direct counts of organisms (bi-
ological)
Time consumption
Faster Slower
Skills required
Less More
Sampling accuracy
Less accurate Better
Confidence with
respect to D2
Lower Higher
According to the Regulation D2, organisms of interest can be divided into 3 size
classes; viable organisms 50 m, viable organisms 10 m and < 50 m and D-2
bacteria (Enterococci, Escherichia coli and Vibrio cholerae) (IMO, 2009). The main
principle for indicative analysis is that it is sufficient to identify the potential compli-
ance using only one indicator group of organisms (Gollasch and David, 2015). In
general, indicative analysis tools for phytoplankton can be practically utilised on-
board with relatively low amount of required ballast water, whereas zooplankton
samples might require further analysis in a laboratory. IMO (2015) has compiled the
general methodological approaches for indicative analysis sampling (Table 4).
Trafi Publications 10-2017
4
Table 4. General methodological approaches for indicative analysis methods testing Regulation D2 compliance
(IMO, 2015).
Organism group
Sampling approach
Notifications
Viable organisms
50 m Visual counts or ste-
reo-microscopy Can be expensive and time-consuming,
requires training
Viable organisms
50 m Visual inspection Most likely limited to only register organ-
isms bigger than 1000 micro-metres
(mm) in minimum dimension
Viable organisms
10 m and < 50
m
Variable fluorometry Only able to monitor photosynthetic phy-
toplankton, underestimating other plank-
tonic organisms in this and other size
fractions
Viable organisms
50 m and 10
m and < 50 m
Photometry, nucleic
acid, ATP, bulk fluo-
rescein diacetate
(FDA), chlorophyll a
Relatively accurate results can be ob-
tained. Some organic compounds can in-
dicate viability for various periods of time
outside the cell, potentially leading to
false results
Viable organisms
50 m and 10
m and < 50 m
Flow cytometry Can be very expensive
Enterococci Fluorometric diagnostic
kit Incubation can be time-consuming
Escherichia coli
Fluorometric diagnostic
kit Incubation can be time-consuming
Vibrio cholerae
(O1 and O139) Test kits Relatively quick indicative tests available
Viable organisms
50 m and 10
m and < 50 m
Pulse counting fluores-
cein diacetate (FDA) Sampling kits most likely larger than the
ones for FDA
1.2 In-tank and in-line sampling for indicative analysis
G2 guidelines of the BWM Convention refer to indicative analyses as prior proce-
dures testing for compliance (IMO, 2015). However, the guidelines do not strictly
describe how the sampling should be conducted from the ballast water tanks or dis-
charge line. As there are various types of ships visiting ports on a daily basis, sam-
pling equipment and accessibility play key roles in obtaining a sample from the bal-
last water (David, 2013).
In-tank sampling has been recommended over in-line sampling by Gollasch and Da-
vid (2015), since it can be conducted before the discharge, whereas in-line sampling
requires discharge, which in turn can risk the destination areas to outbreaks of NIS.
In terms of the alternative sampling points and equipment, water pump sampling via
manholes have resulted in more diverse samples than plankton net samples or sam-
ples from sounding pipes (David, 2013). However, sometimes manholes can be in-
accessible for sampling due to their location or overlaying cargo and the sounding
pipes might be the only option for sampling. In addition, in-tank sampling is only
appropriate when the ballast water treatment has been conducted during the uptake
Trafi Publications 10-2017
5
of the ballast water, since if any part of the treatment process occurs during the dis-
charge, in-tank sampling is not able to measure the success of the treatment system
(IMO, 2009).
According to the representativeness of a sample, in-line sampling is preferred over
in-tank sampling. In-line sampling enables the continuous collection of entrained or-
ganisms for an integrated sample over most of the ballasting/de-ballasting cycle,
whereas the reliability of in-tank sampling can suffer from patchiness of organisms
within the ballast water tanks (Wright, 2012). In general, smaller organisms such as
bacteria tend to have more homogenous distribution than larger organisms such as
zooplankton. Therefore, a representative sample for the smaller organisms (<50 µm)
should consist of an integrated low volume sample, preferably collected over all of
the ballasting/de-ballasting cycle (Gollasch and David, 2010). In-line sampling pro-
vides more reliable and representative samples for compliance monitoring, even
though sampling during de-ballasting prevents further sample collection for detailed
analysis and enables the transfer of NIS if the ship turns out non-compliant. Over-
all, it is stated in the G2 guidelines of the BWM Convention that the samples should
be taken from the discharge line and that sampling through manholes, sounding
pipes or air pipes is not the recommended approach for Regulation D2 compliance
assessments (IMO, 2009). Different sampling options for in-tank and in-line sam-
pling are described in Table 5.
Table 5. Sampling options for in-tank sampling in indicative analyses (David, 2013, Gollasch and David, 2015).
Organism
group
Sampling point
Required
e
quip-
ment
Required
water vo-
lume (litre)
Number of samp-
les
Viable or-
ganisms
50 m
Manhole, sound-
ing pipe or air
vent (In-line)
Plankton net or
pump 300 – 500
(net)
100 (pump)
1 integrated sam-
ple from the whole
water column or
from 3 different
depths (pump)
Viable or-
ganisms
10 m and
< 50 m
Manhole, sound-
ing pipe or air
vent (In-line)
Pump, water col-
umn sampler or
point-source
sampler/bucket
5 – 6 1 integrated sam-
ple from the whole
water column or
from 3 different
depths
Indicator
microbes Manhole, sound-
ing pipe or air
vent (In-line)
Pump, water col-
umn sampler or
point-source
sampler/bucket
1 1 integrated sam-
ple from the whole
water column or
from 3 different
depths
1.3 Recommended principles for the D2 standard compli-
ance testing
According to IMOs’ G2 guidelines, samples are recommended of being concentrated
to a manageable size and the sampling process should be undertaken safely and
practically (IMO, 2009). As relatively longer sampling times can result in underesti-
mation of the present viable organisms, sequential samples of approximately 10
minutes are recommended (David and Gollasch, 2015). Timing of sampling is also
important as the organisms are less likely to be homogenously distributed within the
ballast water tanks. The uneven distribution of organisms in ballast water tanks can
Trafi Publications 10-2017
6
result in potential errors in the compliance determination (Figure 1) (Miller et al.,
2011, Frazier et al., 2013, Costa et al., 2015). In order to prevent under- or overesti-
mation of organism concentrations, sampling time is recommended to avoid taking
sample at the first 5 minutes and the last 5 minutes of the de-ballasting event. The
sampling duration of approximately 10 minutes is also advised to be divided to take
roughly 0,5 litres every minute, instead of taking the entire required sample volume
at once.
As also a single 10 minute sequential sample can under- or overestimate the present
organism concentrations, an average of at least 2 random 10 minute sequential sam-
ples is recommended (David and Gollasch, 2015). Additional attention during sam-
pling should be drawn to ballast water discharge flow rates, as relatively strong flow
rates or sheer forces at the valves nearby sampling points can damage organisms and
lead to falsified results in terms of organism viability. Therefore the valves at the
sampling points should be kept open and the ballast water discharge rate should not
exceed 50 litres per minute.
Size classification of organisms can also become problematic during indicative sam-
pling (David and Gollasch, 2015). The minimum dimension of an organism is de-
signed to include the smallest part of the body and ignore the sizes of spines and an-
tennas (IMO, 2009). However, some indicative sampling devices, such as flow cam-
eras and flow cytometers can misinterpret the organism sizes and therefore divide
organisms into wrong categories, or treat colonies as individuals, specifically for
non-spherical objects (David and Gollasch, 2015). Sizes of the colony forming spe-
cies should be classified as the minimum dimension of an individual, not the colony.
Even though the sampling of only one size category of organisms should be suffi-
cient for the determination of compliance, viable organisms greater or equal to 50
m in minimum dimension tend to exceed the D2 standard limitations more often
than smaller viable organisms and microbes (David, 2013). Additionally, different
size categories of organisms require different sampling approaches. David and Gol-
lasch (2015) and Gollasch et al. (2015) have stated that indicative sampling for or-
ganisms greater or equal to 50 µm and regulation D2 bacteria is not as reliable as
Figure 1. Potential sampling outcomes in compliance determination (Jorgensen et al., 2010,
Frazier et al., 2013).
Trafi Publications 10-2017
7
sampling for organisms smaller than 50 µm and greater or equal to 10 µm. Sampling
methods for larger organisms usually require larger quantities of water for sampling,
as they are present in lower concentrations than smaller organisms and microbes in
the ballast tanks. Detection of bacteria in turn, can be too time-consuming, as Regu-
lation D2 requires the concentration of bacteria in CFUs (Gollasch et al., 2015) and
these bacteria are rarely present even in untreated water (Welschmeyer and Kuo,
2016). Therefore it can be concluded that organisms smaller than 50 µm and greater
or equal to 10 µm is the most reliable indicator group for the compliance of the ship,
and also easiest to prove (David and Gollasch, 2015, Gollasch et al., 2015).
Overall, a Port State Control (PSC) officer should be able to conduct the sampling
with some special training, but without a requirement for academic education in bi-
ology or chemistry (David and Gollasch, 2015). Indicative analysis devices should
also be portable, or at least the samples should be detectable for compliance in a
portable laboratory outside the vessel. Cost-effectiveness of the chosen sampling
methods should also be considered, but as a general advice, relatively expensive but
accurate and appropriate sampling technology is preferred over cheaper and less ac-
curate systems.
2. Materials and methods
The present study was based on existing literature on indicative analysis methods as-
sessing compliance status of ships according to the BWM Convention. The cited lit-
erature consists mainly of scientific studies provided by Google Scholar and Pub-
Med databases, but also of technical and practical reports, publications by govern-
mental and international organizations, training manuals and applicable books and
reviews. No practical sampling or laboratory work was conducted for this study. Lit-
erature on individual indicative methods was searched with the name of the method
in association with ballast water sampling, for example “ATP ballast water sam-
pling”. Relevant citations within key articles were also used to extend the knowledge
around the topic. Overall, 125 references were utilized. In addition, certain key re-
searchers and device manufacturers were contacted to gain further information about
certain evaluated indicative analysis methods.
The studied methods were compared and evaluated based on their feasibility to in-
dicative analysis requirements. As the present study was assigned by the Finnish
Transport Safety Agency, feasibility of each method was assessed to serve their in-
terests. After consultation with the Agency, it was suggested that indicative analysis
sampling devices should cost less than 100,000 €, with a maximum analysis time of
2 hours. Obviously, the fastest devices with the best cost-efficiency and accuracy
were preferred over devices that just fitted in to these frames. Additionally, as de-
vice’s portability was also considered as a key factor, handheld devices were pre-
ferred over portable devices that require larger transportation arrangements.
Trafi Publications 10-2017
8
3. Results
The most well-known indicative sample analysis methods are shown in Table 6, and
they are evaluated in detail in the following chapters. Some methods can analyse
only one size category of organisms, whereas others are able to detect multiple size
categories. For this reason, the analysis methods will be evaluated individually in-
stead of dividing them into different organism size categories
Table 6. Indicative sample analysis methods (Jorgensen et al., 2010, Bradie, 2016, IMO, 2015, David and Gol-
lasch, 2015).
Method Description
Adenosine triphosphate
(ATP)
Measures luminescence in the presence of luciferase enzyme
from seawater extraction. Some test kits give estimation of all bio-
logical contamination through the quantification of bioluminescent
signal coming from the reaction of the Luminase with intracellular
adenosine triphosphate (cATP), the energy carrier of any living
cell.
Fluorescein diacetate (FDA)
staining
FDA: Stains living phytoplankton cells for microscopic inspection.
Pulse counting FDA: Counts fluorescence pulses over specified
threshold from FDA stained organisms, Bulk FDA: Calculates fluo-
rescein production rate and concentration of live cells after incu-
bation, Sytox FDA: stains dead phytoplankton cells.
Pulse amplitude-modulation
(PAM) fluorometry (Also
known as variable fluorome-
try)
Measures photosynthetic activity and phytoplankton biomass
which are considered as indication for viable cells. Analyses living
cells based on variable fluorescence (Fv) of chlorophyll of living
algae.
Microscopy Visual inspection, moving organisms can be counted as viable,
FDA staining: stained cells can be counted as viable.
Flow cytometry Channels samples to the detector and measures stained organ-
isms.
Fluorometric diagnostic kits
(Enterococci & Escherichia
coli)
Detects viable bacteria by testing for the presence of key en-
zymes or nutrient-indicators. Samples are incubated and exam-
ined for fluorescing wells. The number of these wells refers to the
most probable number (MPN) of total bacteria in a sample. Some
of them detect bacteria at 1 cfu/100 ml.
Polymerase chain reaction
(PCR) bacterial RNA detec-
tor
Utilizes designed primers to detect group of genes within bacterial
RNA, induces the multiplication of RNA with integrated fluorescein
and gives a signal in a fluorometer.
3.1 Adenosine triphosphate (ATP) method
The detection of adenosine triphosphate (ATP) as an indicative measure of cellular
biomass has been utilized in the estimation of aquatic organism concentrations al-
ready in the 1970s (Hodson et al., 1976). As a molecule produced by all living or-
ganisms, ATP presents the amount of total living biomass energy measured in a
sample from the ships’ ballast water (Bakalar, 2014). ATP detection is generally
based on bioluminescence originated from the firefly’s (Photinus pyralis) luciferin
or luciferase complexes (van Slooten et al., 2015). ATP is generally extracted from a
sample and reacted with luciferin/luciferase (Karl, 1993). As a measure of metabolic
activity, ATP is acknowledged as a relatively good viability indicator for unicellular
Trafi Publications 10-2017
9
organisms with a tendency for asexual reproduction, even though metabolic activity
does not ensure viability (van Slooten et al., 2015).
3.1.1 Sampling approach
The ATP method contains 3 distinctive steps; sample concentration, ATP extraction
and determination utilizing a sample swab with liquid-stable reagents and a
handheld luminometer (First and Drake, 2014). A water sample is collected and run
through a cartridge with a membrane filter using a syringe filter for organism con-
centration. This step also purifies the sample by removing dissolved compounds and
residual oxidants from interfering with the luminogenic reaction. ATP extraction in-
cludes the application of lysis buffers through the filter to lyse cells and extract ATP.
The extraction can be diluted if necessary. The dilution is pipetted to the sample
swab for the luminometer analysis, which detects relative luminescence units
(RLUs). In order to determine the size categories of organism concentrations, water
samples can be pre-filtered to remove larger particles (van Slooten et al., 2015). An
example from the sampling approach for an ATP analysis is illustrated in Figure
2.The entire sampling protocol for this method is provided in Appendix A.
Figure 2. Sampling approach for the ATP analysis. Redrawn from SGS (2015).
Trafi Publications 10-2017
10
3.1.2 Feasibility
ATP analysis can be an effective indicator for compliance determination, since it
provides the basis for the determination of total viable plankton biomass (Waite et
al., 2003). The analysis method enables the evaluation of all organism size catego-
ries and the sampling devices are relatively affordable (<10,000$) without substan-
tial running costs (van Slooten et al., 2015, Bradie, 2016). As long as the ATP rea-
gents can be contained in room temperature, water quality characteristics of the sam-
ples, such as low salinity and high turbidity should not influence the outcome of
ATP analysis (van Slooten et al., 2015). However, highly saline ballast water in turn,
may require additional steps of dilution for the sampling protocol. Relatively high
concentrations of total suspended solids (TSS) can however reduce the efficiency of
filters (First et al., 2014). Sample contamination risks using the ATP analysis
method, such as leaking or damaged filters can be considered as relatively low, alt-
hough some delays in the reduction of ATP levels can occur after UV treatment as
the affected cells do not die immediately after treatment (van Slooten et al., 2015).
The sampling time for ATP analysis is also reasonable, as results can be obtained in
less than an hour (SGS, 2015, Bradie, 2016).
However, ATP analyses can be somewhat problematic. Presence of dissolved metals
in the water samples can cause underestimations in detected biomass by inhibiting
the light production (Sudhaharan and Reddy, 2000). Azam and Hodson (1977) in
turn observed that the presence of free ATP in the environment will result in overes-
timation of the biomass. The effects of dissolved compounds can be eliminated by
applying a filter membrane (First and Drake, 2013), or ATP extraction with boiling
tromethamine or activated carbon, although the mentioned extraction techniques are
considered being too time-consuming and complex for PSC officers (van Slooten et
al., 2015). Sample filtration can be considered highly important, since otherwise it is
not possible to differentiate the sampled organism sizes (Bakalar, 2014). Addition-
ally, UV treatment for ballast water has been detected to increase the amount of cel-
lular ATP in bacteria (Villaverde et al., 1986), which complicates interpretation of
the results from UV-radiated samples (First and Drake, 2013).
ATP analyses have been studied in laboratories, as well as during onboard experi-
ments, and they have shown relatively interesting and also contrasting results.
Wright et al. (2015) studied ATP sampling on organisms above 10 µm in minimum
dimension before and after filtration and UV radiation treatment. The ATP analysis
in this study revealed great results with relatively short extraction time, as the ATP
concentration decreased by 99 % after the (UV+filtration) treatment. Even though
the extraction took only 5 minutes, the samples were frozen and analysed further in a
laboratory. Therefore it remains unclear, whether the same results would have been
obtained in situ. Van Slooten et al. (2015) also reported a strong decrease in ATP
levels after UV treatment in a laboratory experiment. In addition, they experienced
that the ATP sampling method with syringes and filters was relatively easy to use,
portable in a light briefcase and not excessively time-consuming.
First and Drake (2014), in turn, studied the efficiency of UV and chlorine dioxide
treatments using ATP analysis on collected seawater samples in a laboratory. The
ATP levels did not decline significantly after the UV treatment in this study, even
though indicative Pulse Amplitude-modulation (PAM) fluorometry method detected
significant decreases after the treatment. However, ATP analysis detected significant
decreases after chlorine dioxide treatment, suggesting that appropriate ballast water
Trafi Publications 10-2017
11
verification methods can also be treatment-specific (First and Drake, 2014). De
Lafontaine et al. (2009) noted that ATP analyses can be useful in measuring growth
dynamics of viable yeast populations in individual experiments, but also concluded
that ATP measurements did not response accurately to yeast cell density in all cases.
They used ATP analysis while investigating the efficiency of yeast-based deoxygen-
ation treatment method for cold water conditions in a laboratory environment.
Anyhow, ATP analysis has shown promising results after high voltage electricity
treatment (Hwang et al., 2010). This laboratory experiment revealed that the ATP
levels of zooplankton, phytoplankton and bacteria, all decreased significantly after
the electric pulse treatment. Another laboratory experiment by Penru et al. (2012)
found significant declines in cellular ATP from seawater samples after UV treat-
ment. In fact, one of the most recent studies in ATP method testing for compliance
(Welschmeyer and Kuo, 2016) reviewed that ATP devices can be used on-board as
portable devices, provide results quickly, indicate results efficiently regardless of the
treatment method and can be calibrated to correspond only certain organism size cat-
egory.
3.2 Fluorescein diacetate (FDA) method
Fluorescein diacetate (FDA) represents a cell-specific identification method for cel-
lular viability (Rotman and Papermaster, 1966). The method estimates living plank-
tonic biomass from quantitative and enzymatic transformation of FDA into fluores-
cein, its fluorescent product (Welschmeyer and Maurer, 2011). Applications of the
FDA method can have some variation, as the fluorescent signal can be inconsistent
and FDA does not necessarily stain all organisms (Dorsey et al., 1989, Garvey et al.,
2007). Additionally, the permeability of FDA throughout the cell membrane has re-
sulted in the development of derivatives (First and Drake, 2013), such as 5-chloro-
methylfluorescein diacetate (CMFDA) and carboxyfluorescein diacetate (CFDA),
which have the advantage of better cellular retention through their reactive proper-
ties in comparison to FDA (Steinberg et al., 2011, Gorokhova et al., 2012).
Another approach for staining is to identify dead cells with nucleic acid stains im-
permeable to intact cell membranes, such as the fluorophore SYTOX green and TO-
PRO-1 iodide (Gorokhova et al., 2012), and detect the concentration of living organ-
isms by subtraction (First and Drake, 2013). Detection of cytoplasmic membrane in-
tegrity utilizes probes that fluorescent only when bound to nucleic acids (Berges and
Falkowski, 1998). In general, further tests are required in the detection of the total
number of dead cells within a sample (First and Drake, 2013). An additional method
using FDA staining is FDA pulse counting, which in turn utilizes a detecting ana-
lyser to count fluorescence pulses from stained viable organisms (Nakata et al.,
2014). The analyser estimates the viable organism concentration from the pulses
with a practical threshold.
3.2.1 Sampling approach
Even though fluorescein is a non-fluorescein compound, it can be utilized to emit
fluorescein through hydrolysis by biological enzyme activity (Rotman and Paper-
master, 1966). Green fluorescent emission originates from FDA associated with es-
terase activity present in all living organisms (Figure 3) (Welschmeyer and Maurer,
Trafi Publications 10-2017
12
2011). FDA analyses can be used to quantify numeric counts of viable cells in asso-
ciation with an epifluorescent microscope or flow cytometry, or alternatively, as a
bulk indicator for viable cell biomass.
Figure 3. Methodological approach for FDA as a marker for viable organisms (Welschmeyer and Maurer, 2011).
FDA analysis includes staining, incubation and counting of samples (Steinberg et
al., 2011). A proportion of a sample is mixed with FDA, CFDA or CMFDA into mi-
crofuge tubes for sample staining. The incubation step is conducted in the dark at
room temperature for 10 – 30 minutes, depending on the specific FDA method used
(Steinberg et al., 2011, Gorokhova et al., 2012, van Slooten et al., 2015). The organ-
isms in incubated samples can be analysed with an epifluorescent microscope as flu-
orescing and non-fluorescing, or mobile and immobile (Steinberg et al., 2011). Al-
ternatively, the fluorescence can be analysed with a fluorometer (van Slooten et al.,
2015).
As mentioned earlier, staining of dead cells in a sample requires further methods to
transform this outcome into viable organism concentrations (First and Drake, 2013).
For example, DNA probes permeable to both, dead and living cells can be used to
counter-label the cells, of which the membrane permeable-labels can be determined
using flow cytometry, epifluorescent microscope or fluorescence measuring plate
readers (Peeters et al., 2008, Peperzak and Brussaard, 2011, Steinberg et al., 2012).
FDA Bulk analysis is generally based on a fluorometric analysis of the extracellular
bulk liquid containing the suspended cellular material (Welschmeyer and Maurer,
2011). FDA bulk analysis requires filtration of a sample, applying the filter into a
cuvette containing reagent buffer and mixing with FDA reagent (Maurer, 2013). The
incubation time for bulk FDA analysis is 1 hour, after which the sample is squeezed
from the filter into a centrifuge, spinned down and the fluorescence is measured with
a spectrofluorometer.
Pulse counting FDA analysis detects fluorescence pulses and estimates the organ-
isms viability based on the strength of the pulse (Figure 4) (Nakata et al., 2014). The
analysis method requires a mixture of FDA and a sample, which is thereafter run
through the pulse counting device. The device counts the organism concentration
Trafi Publications 10-2017
13
from the sample and shows it on the screen panel on the device. The operability of
FDA pulse counting device is represented in Appendix B.
3.2.2 Feasibility
Various applications of FDA analyses have been widely acknowledged as appropri-
ate methods to measure organism viability in numerous environmental samples
(Adam and Duncan, 2000 described in Akram et al., 2015). FDA analysis enables
the sampling of all organism size categories and depending on the elected method,
the sampling duration varies between 30 minutes and a few hours and is thus reason-
able (Welschmeyer and Maurer, 2011, Bradie, 2016). Bakalar (2014) also estab-
lished that FDA devices can be affordably priced, starting from 450 US dollars with
minimal running costs. However, the risk of filter damage and leakage is present in
FDA analyses that require the use of filters (van Slooten et al., 2015). Additionally,
water turbidity greater than 20 NTU can decrease the accuracy of fluorescence de-
tection (Bradie, 2016). Granqvist and Mattila (2004) and Engström-Öst and Mattila
(2008) reported that water turbidity in the Baltic Sea region can vary between 0.5
and 45 NTU. Anyhow, Waite et al. (2003) reported that generally the most turbid
waters in natural conditions range between 10 and 15 NTU.
FDA analysis can however produce inconsistent fluorescence signal and FDA has
been observed not being able to stain all organisms (Garvey et al., 2007). Adams et
al. (2014) and MacIntyre and Cullen (2016) attempted the utilization of combined
FDA and CMFDA stains in organism viability assessment, as derivatives of FDA
are impermeable across the cell membrane (First and Drake, 2013). According to
their suggestion, even FDA+CMFDA analyses cannot be considered as sufficient vi-
ability assessment methods due to relatively significant risk of overestimation of live
organisms when fluorescing dead organisms. The overestimation problem has also
been reported in association with CFDA analysis on UV-treated seawater, as the cell
membrane can be intact within the UV-treated cells even some days after the treat-
ment (Tobiesen et al., 2011). Steinberg et al. (2011) reported similar problems with
the combined FDA and CMFDA stains assessing heterotrophic and mixotrophic di-
noflagellates, but in contrast found the same method successful in viability assess-
ment of protists. In general, assessments utilizing any type of staining method have
the issue of not being able to make a distinction between living and viable cells
Figure 4. Methodological approach for pulse counting FDA (Nakata
et al., 2014).
Trafi Publications 10-2017
14
(Reavie et al., 2010) and the efficiency of FDA analyses can also depend on the bal-
last water treatment method (Akram et al., 2015).
A laboratory experiment by van Slooten et al. (2015) revealed that the conducted
FDA analysis was proven both, more time-consuming and less accurate than the
conducted ATP analysis. Excessive time consumption of various FDA analyses has
been reported also in other studies. On top of the longer sampling duration, Stein-
berg et al. (2012) and Carney et al. (2013) reported that fluorescence-based tech-
niques tend to struggle with lower sampling volumes and uneven distribution of
cells on membrane filters or in counting chambers. In addition, staining of dead cells
with nucleic acid stains includes not only the staining of dead cells, but also the re-
staining of both, living and dead cells for the subtraction (Steinberg et al., 2012).
The process is inevitably vulnerable for measurement errors, as it requires a second-
ary count for the cells (First and Drake, 2013) and underestimations of non-viable
cells can occur due to viable labelling efficiency of cells with damaged DNA (Leb-
aron et al., 1998).
As indicative compliance testing requires quick and portable testing devices, some
efforts and improvements have been also applied to FDA analyses. Welschmeyer
and Maurer (2011) developed a quicker and more simplified bulk FDA analysis
method by combining a series of analysis steps. Even though these sampling proce-
dures can be conducted within 30 minutes for one sample, the method was also
tested on a full-scale onboard experiment revealing that the completion of the proce-
dures took 3 hours, despite leading to relatively promising results on successful bal-
last water treatment. Nakata et al. (2014) conducted a pulse counting FDA experi-
ment with a portable sampling device obtaining results in 10-30 minutes. They com-
pared the measured FDA pulses to organisms counted with a microscope. Even
though the study found a highly significant correlation between the counts and
pulses without disturbing background fluorescence, the reliability of this device suf-
fers from lack of additional studies and underestimation of organisms (Bradie,
2016).
Overall, it has been widely noted that the method of staining dead or living cells ap-
pear to have issues with false positive outcomes (Tobiesen et al., 2011, Zetsche and
Meysman, 2012, Adams et al., 2014, Wright and Welschmeyer, 2015). Majority of
FDA analyses also suffer from relatively long incubation periods and require further
onboard studies to establish the approach as a reliable indicative analysis method.
3.3 Pulse amplitude-modulation (PAM) method
PAM fluorometry measures the photochemical efficiency of photosystem II (PS II),
fluorometric character of a particle in phytoplankton containing origins of chloro-
phyll fluorescence (Schreiber, 2004, Bakalar, 2014). Therefore the methodology in-
dicates the viability of phytoplankton regardless of their size, additionally enabling a
qualitative or quantitative indication of phytoplankton community through photo-
synthetic activity (Schreiber et al., 2007, Gollasch et al., 2015). PAM analyses can
be conducted as bulk analysis or a measurement for single cell counts (Villareal,
2004). Somewhat similarly to FDA analysis, PAM method measures selective in-
creasing fluorescence signals based on transmitting short, but intensive light pulses
(Bakalar, 2014).
Trafi Publications 10-2017
15
3.3.1 Sampling approach
PAM fluorometry refers to the detection of photosynthetic performance parameters
utilizing fluorescing saturation pulses to stimulate species sensitive to light (Figure
5) (Heraud and Beardall, 2000). The methodology measures quantum yield of fluo-
rescence also known as instantaneous variable or maximum fluorescence, based on
minimum (F0) and maximum (FM) fluorescence yields, and photochemical yield (Y),
calculated from the following equation (Genty et al., 1989, First and Drake, 2014,
Wier et al., 2015);
= 1000×
 
According to this equation where “Y” indicates the photosynthetic efficiency, the
values of Y for living and healthy marine phytoplankton generally range between
400 and 600 (First and Drake, 2013, Wier et al., 2015). In addition, low values of Y
are an indication of dysfunctional photosynthesis. As the quantum yield of the asso-
ciated electron transport is directly proportional to the output of photochemical fluo-
rescence quenching (Genty et al., 1989), photosynthetic efficiency can be also deter-
mined by using an equation that excludes the multiplication with 1000 (Genty et al.,
1989, Bradie, 2016). In this occasion, the typical values of instantaneous fluores-
cence range between 0 and 0.8, and values above 0.3 represent viable phytoplankton
populations (Villareal, 2004, Stehouwer et al., 2010, First and Lake, 2014, Bradie,
2016).
Figure 5. Typical fluorescence output from a PAM, including device being switched on (LL),
activation with 10 second actinic light pulses (AL). SF (Saturation Flashes) represents re-
sponses to AL, whereas Fm’ is fluorescence maxima (White and Cricthley, 1999).
Trafi Publications 10-2017
16
In order to obtain a measurement for the minimum fluorescence (F0), the samples
generally need to be pre-adapted to dark. Dark-adaptation period of 10 minutes is
considered sufficient (Heraud and Beardall, 2000), but the duration of the dark-adap-
tation can vary widely between 10 seconds and 30 minutes depending on the sam-
pling device and applied sampling protocol (Heraud and Beardall, 2000, Villareal,
2004, Martinez et al., 2012, First and Drake, 2014, Bradie, 2016). Alternatively,
dark-adaptation can be excluded completely to minimize the effect of potential con-
tainment artifacts (Villareal, 2004).
The sampling stages include discretionary syringe filtration and adaptation to dark,
sample placement into a cuvette and yield measurement by exposing the sample to
short sequences of increasing actinic light pulses (Villareal, 2004, Bradie, 2016).
The measuring device is usually operating with a cell-counting flow camera or an
epifluorescent microscope (Villareal, 2004, Gollasch, 2012). The detection of organ-
ism concentration generally requires a conversion from the recorded photochemical
yields into cell concentrations, depending on the applied device (Bradie, 2016).
3.3.2 Feasibility
David and Gollasch (2015) concluded that according to their knowledge, PAM fluo-
rometry is the best indicative analysis sampling tool with the advantages of quick-
ness, portability and ease of use. PAM devices enable the viability results being ob-
tained within a matter of minutes, some of them even faster (First and Drake, 2013,
Bakalar, 2014, van Slooten et al., 2015). The price range for PAM sampling devices
varies roughly between 4,000 and 15,000 US$ with relatively minimal running costs
(Bradie, 2016). PAM fluorometers are mainly designed to indicate the presence of
phytoplankton in the < 50µm and 10µm size category (Gollasch et al., 2015), but
Bradie (2016) reported that some devices are also able to detect organism concentra-
tions from the 50 µm category. Additionally Bradie (2016) noted that the accuracy
of some PAM fluorometers can decrease at water turbidity greater than 20 NTU.
Due to the simplicity and automated nature of PAM sampling technique, the risk for
sample contamination using this method can be considered relatively minimal (Ba-
kalar, 2014, van Slooten et al., 2015).
Several studies, such as First and Drake (2014), Gollasch et al. (2015), van Slooten
et al. (2015) and Casas-Monroy et al. (2016) have also detected limitations while
sampling with PAM devices. Understanding of PAM sampling technology is essen-
tial, as the methodology is sensitive to dead organisms and vulnerable to systematic
errors with treatment systems that eliminate the organisms but do not necessarily re-
move the remaining chlorophyll a (First and Drake, 2014). As Casas-Monroy et al.
(2016) stated PAM devices are vulnerable to under- or overestimation of certain or-
ganism groups. Therefore, it is recommended that additional tests for different or-
ganism groups are included along with PAM measurements to confirm the compli-
ance status (Gollasch et al., 2015).
Probably one of the most significant limitations associated with PAM fluorometry is
that it measures only concentrations of autotrophic organisms via chlorophyll a (van
Slooten et al., 2015), whereas the BWM Convention requires the determination of
all organisms in the targeted size class (Gollasch et al., 2015). Autotrophs refer to
organisms that are able to utilize inorganic materials as a source of energy via photo-
synthesis and chemosynthesis (mainly phytoplankton and certain bacteria) (van
Slooten et al., 2015). On the contrary, heterotrophs (e.g. ciliates and protozoa) utilize
Trafi Publications 10-2017
17
organic compounds as a principal source of food and their presence cannot be de-
tected using chlorophyll-based methods. In addition, methods testing for the effi-
ciency of PS II cannot detect the presence of cyanobacteria, of which fluorescence
derives from other pigments than of chlorophyll’s (Sugget et al., 2006).
Despite the limitations of PAM devices, they have been evaluated as one of the most
appropriate indicative sampling analysis methods (van Slooten et al., 2015, Casas-
Monroy et al., 2016). Shannon et al. (2009) conducted laboratory experiments on
treated and untreated artificial seawater samples, successfully detecting variable flu-
orescence yield with PAM fluorometry in comparison to decreasing photosynthetic
efficiency, altered with photosynthesis inhibiting herbicide 3-(3,4-dichloro-phenyl)-
1,1-dimetylurea (DCMU). After the execution of laboratory experiments by van
Slooten et al. (2015), Gollasch et al. (2015) and Casas-Monroy et al. (2016), PAM
fluorometry has been found superior over ATP and FDA methods in terms of sam-
pling duration, ease of use, device portability and occasionally even sampling accu-
racy.
Studies by Gollasch et al. (2015) and Bradie (2016) have additionally compared the
differences between several PAM devices produced by different manufacturers.
BBE 10 cells, Hach BW680, Turner Designs Ballast-Check and Walz Water PAM
were all relatively fast, portable and required minimal amount of training without
significant differences in these categories. However, Hach BW680 device provided
the most consistent results between replicates during the study by Bradie (2016), and
was also found easiest to use as a handheld device without any filtration steps (Gol-
lasch et al., 2015). Bradie (2016) also revealed that out of these PAM devices, Hach
BW680 was the cheapest to purchase.
3.4 Microscopy methods
Microscopic viability analysis for zooplankton can be conducted by stimulating or-
ganisms through poking them and detecting motility (David and Gollasch, 2015).
The main principle is that moving organisms can be counted as viable. Alternatively,
microscopic analysis for phytoplankton can be conducted with FDA staining since
movement cannot be considered as a reliable indicator for viability of autotrophs
(Bradie, 2016). Staining for epifluorescent microscope analysis can also be con-
ducted with CMFDA or Sytox (David and Gollasch, 2015).
3.4.1 Sampling approach and feasibility
Microscopy method is considered as a standard method for ballast water samples,
analysing organisms larger or equal to 50 µm or organisms smaller than 50 µm and
larger or equal to 10 µm in minimum dimension (Bradie, 2016). The microscopy
method approach includes concentration of samples from larger volumes of water
into concentrated samples, mixing of samples and pipetting samples for the micro-
scopical inspection. In terms of feasibility, microscopy analyses have the advantage
of directly referring results to organisms per volume (Wright et al., 2015). However,
microscopic inspection requires usually prolonged periods of time, as well as spe-
cialized knowledge to identify viable organisms (First and Drake, 2013, Wright et
al., 2015). The inspection can take several hours and assessing motility refers to liv-
ing organisms, instead of confirming reproductive ability of these organisms, in
other words viability (Bradie, 2016).
Trafi Publications 10-2017
18
The sampling approach for epifluorescent microscope analysis utilizing FDA,
CFDA, CMFDA or Sytox stains has been described earlier in section 3.2.1. Any-
how, previous studies on epifluorescent microscopy, such as First and Drake (2013),
Gollasch et al. (2015), Casas-Monroy et al. (2016) and First et al. (2016) indicate
that the approach utilized for direct counts can take several hours and requires not
only biological expertise, but also laboratory surroundings (Gollasch et al., 2015). In
addition, as cysts do not move or respond to fluorescent stains, microscope tech-
niques are most likely unable to detect encysting organisms in a simple or efficient
manner (First et al., 2016).
3.5 Flow cytometry method
Flow cytometry is designed to detect the abundance of phytoplankton and bacteria
(Gasol and Del Giorgio, 2000, Veldhuis and Kraay, 2000). The methodology utilizes
laser beam, which stimulates cells to scatter light when encountering laser and emit
fluorescent light after excitation by the laser (Brussaard et al., 2000). Flow cytome-
try detects the amount of phytoplankton cells based on red fluorescence signal (>610
nm) originating from chlorophyll a (Veldhuis and Kraay, 2000, Veldhuis et al.,
2001), whereas the abundance of bacteria can be determined with additional DNA-
specific fluorescent staining (Li et al., 1995, Gasol and Del Giorgio, 2000).
3.5.1 Sampling approach
Flow cytometry is an efficient technology for the analysis of individual cell concen-
trations from heterogeneous populations (Picot et al., 2012). It is used to enumerate
and characterize cells from multicellular organisms and single-celled microbes
(Shapiro, 1983, Olsen et al., 2015). Flow cytometer measures scattered light at vari-
ous angles and fluorescence emission, as the cells flow through one or several laser
beams (Picot et al., 2012). Flow cytometry analysis includes the hydrodynamic fo-
cusing for cellular suspension by the fluidic system, detection of excitation source
and fluorescence emission by optical collection system as the cells interact with the
laser beam, and digitalization of the signal by the electronic system for computer
analysis (Figure 6). Forward scatter light (FSC) refers to the size of the cell, side
scatter light (SSC) relates to the structure and shape of the cell, whereas specific flu-
orescence emission indicates the cell characteristics.
Trafi Publications 10-2017
19
Figure 6. Flow cytometry analysis approach (Picot et al., 2012).
Flow cytometry assessment can be conducted and utilized in a number of ways (Fig-
ure 7) (Lomas et al., 2011). Flow cytometers can be automated devices that include
only the appliance of the water sample (Bakalar, 2014), or alternatively, methodo-
logical steps like staining and incubation can be conducted by the examiner (Veld-
huis and Kraay, 2000, Joachimsthal et al., 2003, Binet and Stauber, 2006). The addi-
tional steps include for example viability assessment for phytoplankton utilizing 14C
incubation method (Veldhuis et al., 2006), cell counting for bulk FDA samples
(Bradie, 2016), direct cell-specific fluorescent analysis (Peperzak and Brussaard,
2011) or utilization of staining methods for DNA content and viability determination
(Veldhuis et al., 2001, Lomas et al., 2011).
Trafi Publications 10-2017
20
Figure 7. Potential variation in flow cytometric assessments (Lomas et al., 2011).
3.5.2 Feasibility
As the purpose of this literature review is to evaluate the most suitable and practical
sampling devices for indicative ballast water analyses, only automated and simple
flow cytometers will be discussed in this section. Flow cytometry assessment is able
to provide an automated, quick and accurate method to sample plankton communi-
ties and bacteria (Joachimsthal et al., 2003, Veldhuis et al., 2006, Bakalar, 2011).
Bakalar (2014) added that flow cytometry provided the most accurate results in
comparison to FDA, ATP, PAM and automated colorimetry analyses in a multidi-
mensional projection ranking analysis. Some automated flow cytometry devices
Trafi Publications 10-2017
21
have the advantage of being able to provide results in a few minutes (Stehouwer et
al., 2013).
Some studies have represented estimations that flow cytometric devices can be rela-
tively expensive, ranging approximately between 18,000$ and 200,000€ (Stehouwer
et al., 2013, Bakalar, 2014). Comprehensive understanding of the applied flow cy-
tometer methodology is essential, as several studies, such as Binet and Stauber
(2006), Peperzak and Brussaard (2011), Olsen et al. (2015), tend to highlight quick-
ness of automated flow cytometer as the advantage of the technology, even though
most sampling approaches associated with these studies include sample preparation,
incubation or staining procedures that can take significant amount of time up to few
days. Flow cytometric analysis has also expressed similar limitations to FDA anal-
yses by not being able to make a distinction between dead and living cells (Reavie et
al., 2010, Bradie, 2016). Reavie et al. (2010) concluded that flow cytometry cannot
be considered as an automated substitute for microscopy for this reason. Olsen et al.
(2015) in turn concluded that flow cytometry can differentiate live and dead cells for
certain species, but struggle making a distinction between reproductively viable and
non-viable cells.
D2 standard of the BWM Convention does not include the detection of organisms
<10µm in minimum dimension, even though 90% of all phytoplankton organisms
can fall into this category (Olenina et al., 2006, van der Star et al., 2011). This size
category additionally includes other organisms, such as micro-zooplankton and pro-
tozoa. If this category is to be assessed in the future, flow cytometry can provide a
reasonable sampling approach for smaller organisms, as they are usually present in
relatively high concentrations.
3.6 Bacterial identification methods
There are several methodological approaches for the identification of bacteria in bal-
last water (Gollasch et al., 2012). It is important to emphasize that all methods that
measure bacterial abundance in colony forming units (CFU) require a minimum of 4
hours incubation time for the samples, and the D2 standard requirement for bacteria
is also expressed as CFUs (IMO, 2009). The following chapters are designed to
evaluate the sampling approach and feasibility for each bacterial sampling method,
including a few methods that do not measure bacteria in CFUs.
3.6.1 Detection of bacterial enzymes
Bacterial enzyme detection devices are fluorometers that aim to trace bacteria-spe-
cific detection enzymes by producing fluorescence from additions of key fluoro-
genic substrates, which are hydrolysed by the enzyme (Gollasch et al., 2012). Such
devices enable the identification of certain bacteria, including E. coli and Entero-
cocci within a matter minutes. They are unsuitable for measuring CFUs, as they can
only detect the presence or absence of the targeted bacteria.
Use of most handheld fluorometers requires minimal skills as the sampling approach
includes only the addition of a reagent/substrate to the water sample (Gollasch et al.,
2012). The device expresses the presence or absence of bacteria on a screen associ-
ated with the device. Fluorometers are able to record the readings in approximately
10 to 20 minutes without any incubation. Even though the technology is unable to
Trafi Publications 10-2017
22
represent results as CFUs, semi-quantitative results can be obtained on whether the
sample contains high or low amount of bacterial enzymes. Familiarization to de-
tailed device descriptions is advised, as they may require appliance of chemicals and
incubation, and the sampling duration for some portable fluorometers can be signifi-
cantly longer (Gollasch et al., 2012).
The most probable number (MPN) methodology contains also variations within the
method, but the basic idea of the MPN approach is to detect the most probable num-
ber of CFUs per 100 ml of E. coli or Enterococci based on number of positively flu-
orescing wells (Budnick et al., 1996, Weisberg et al., 2007, Maranda et al., 2013).
Enterolert and colilert test kits by IDEXX laboratories Inc. (Westbrook, Maine) uti-
lize fluorescing nutrient indicator substrates after the metabolization by the targeted
bacteria (Budnick et al., 1996, Cangelosi, 2011). The method includes an incubation
time between 18 and 24 hours (Weisberg et al., 2007, Bradie, 2016). Alternatively,
IDEXX has also developed a method to quantify heterotrophic plate counts (HPC) in
water, which can be altered to MPNs with a 48 hour incubation time (Bradie, 2016).
The MPN method has been found more efficient, effortless and accurate than tradi-
tional bacteria detection method using membrane filters (Budnick et al., 1996). Even
though MPN techniques are the basis of bacterial assessments, their completion re-
quires several hours and can be therefore considered unsuitable of being conducted
within the turnaround time of a ship in ports (Wright et al., 2015).
3.6.2 Real-time Polymerase chain reaction (PCR)
An alternative approach for the detection of bacterial concentration within a water
sample is through genetic nucleic acid priming for E. coli, Enterococci and V. chol-
erae (Weisberg et al., 2007, Fykse et al., 2012). PCR is designed to utilize various
reagents and schemes of temperature alterations to anneal and denature sequences of
nucleic acid for exponential amplification of the targeted genes (Saiki et al., 1985,
Weisberg et al., 2007). Real-time PCR is used to amplify DNA sequences, whereas
real-time NASBA (Nucleic Acid Sequence Based Amplification) is used to amplify
RNA sequences (Weisberg et al., 2007, Fykse et al., 2012).
The method steps include sample filtration, DNA or RNA extraction and the real-
time PCR or NASBA. The analysis can be conducted within 7 hours for real-time
PCR and within 9 hours for real-time NASBA and they describe the results in 1
CFU/100 ml of V. cholerae (Fykse et al., 2012). The detailed analysis protocol is de-
scribed in Fykse et al. (2012). Additionally, quantitative PCR method for the pres-
ence of Enterococci have been able to conduct within 2 – 3 hours for recreational
waters (Haugland et al., 2005), but it remains unclear whether this methodology
could be adjusted to ballast water compliance analyses.
3.6.3 Colorimetric bacteria detection
Colorimetric bacteria detectors are also designed for fast indicative analysis detect-
ing the presence or absence of the targeted bacteria without an estimation of the
amount of CFUs (Gollasch et al., 2012). The test kits are using known antibodies for
the detection of E. coli and V. cholerae (Gollasch et al., 2012), and they can be usu-
ally analysed within 15 minutes (Bakalar, 2014).
Trafi Publications 10-2017
23
The sampling procedures include mixing a few drops of a sample into a test tube
containing reagent, transferring the sample into a swab, placing the swab into the
sampling device and adding the chase buffer (Gollasch et al., 2012). The test takes
approximately 15 minutes to create a distinct colour reaction for positive samples.
There are also other similar colorimetric bacteria detection packages that can be
done in 15 – 30 minutes and additionally some test kits have further developed a
more detailed method with a 6 to 24 hour incubation time to receive more accurate
results. This methodology requires further testing to confirm its operability in terms
of varying salinities and defining the bacteria detection limits (Gollasch et al., 2012).
3.6.4 Other methods for detection of bacteria
As one of the objectives of the present study was to represent all the existing indica-
tive analysis methods for compliance with Regulation D2, the following methods
will be featured even though they have been applied to indicative sampling strate-
gies relatively sparsely. The sampling approach and feasibility of these methods was
evaluated to the most appropriate extent.
The staining of dead and live cells using Sytox and FDA stains for analysis of bacte-
rial concentration with flow cytometry has been evaluated earlier under section 3.5.
The detection for the concentrations of E. coli and Enterococci can also be done us-
ing petri dishes and films (Gollasch et al., 2012). Using petri dishes, the procedures
can be conducted by filtering the sample, incubating for 48 hours and then visually
assessing the CFUs from the discs. Utilization of petri films requires pipetting 1 ml
of the sample into a plate and sealing the plate with a lid. Even though first results
can be visually detected after 6 hours, the test for total bacteria count takes 2 to 3
days, whereas the number of colony forming coliform bacteria can be observed after
incubation of 1 day. Each petri film product is designed for specific bacteria, includ-
ing E. coli and Enterococci petri films (Gollasch et al., 2012).
Alternatively, bacteria can be sampled utilizing microcalorimetry, respirometer or
detection of ribosomal ribonucleic acids (rRNA) (Wadsö, 2002, Gollasch et al.,
2012, First and Drake, 2013, Bradie, 2016). Isothermal microcalorimeters are able to
detect heat generation through metabolism by holding stable sample temperature and
measuring heat generation from biological processes indicating the activity or con-
centration of organisms (Johnson et al., 2009, Braissant et al., 2010, First and Drake,
2013). The targeted organism concentration and activity is related to the time re-
quired to detect metabolic activity. Even though technological development has in-
creased the cost-efficiency and portability of microcalorimeters (Braissant et al.,
2010), this method is yet to be tested comprehensively on ballast water applications.
Similarly to microcalorimeter, Bactest have designed a portable Speedy Breedy pre-
cision respirometer that detects microbial respiration through pressure transients,
which refers to gaseous exchanges (Bradie, 2016). A sample is added into a 50 ml
closed culture vessel including predetermined nutrient medium, resulting in the
growth of micro-organisms. The method is suitable for E. coli and Enterococci and
results can be obtained within 12 hours. Hybriscan rRNA detection method in turn
can be utilized to analyse the presence of E. coli rRNA with species-specific probes
through events of hybridization (Gollasch et al., 2012). The method aims to detect
only living cells and allows the determination of non-culturable microorganisms.
Trafi Publications 10-2017
24
Results can be obtained in approximately 12 to 26 hours, but further research is sug-
gested to evaluate whether the method can be applied to the detection of colony
forming bacteria.
3.7 Other potential indicative methods
Equally to the previous section (3.6.4), the purpose of this chapter is to represent al-
ternative indicative analysis methods that have been discussed in the literature but
are not among the most studied, and evaluate their sampling approaches and feasibil-
ity. Some methods have similarities or variations to aforementioned methods,
whereas some bring completely new approaches to indicative analysis techniques.
Additionally, Gollasch et al. (2012) and First and Drake (2013) have briefly pre-
sented further methods, such as variable oxygen measurements that can be assessed
to test Regulation D2 compliance.
3.7.1 Microfluidic lab-on-chip biosensor
Time-consumption, series of analysis steps required and poor portability of PCR and
microarray approaches has led to the development of portable, microfluidic lab-on-
chip devices (Vilkner et al., 2004, Senapati et al., 2009). The methodology has the
potential to improve reaction efficiency, reduce sampling duration, simplify sam-
pling protocol and decrease the quantity of loading samples (Senapati et al., 2009).
The approach is based on genetic detection assay with a bead-based microfluidic
platform, where probe-operated beads detect fluorescence-tagged target DNA. Alter-
natively, lab-on-chip devices can utilize resistive pulse sensors and a laser to detect
chlorophyll fluorescence (Song et al., 2012, Wang et al., 2013, Song et al., 2014).
The sampling steps for bead-based microfluidic detection platform include fabrica-
tion of microchip using glass slides, filter fabrication within the microchannel, oper-
ating the probe on beads, asymmetric PCR and hybridization assay (Senapati et al.,
2009). Detailed sampling protocol is described in Senapati et al. (2009). The total
detection time is 1 hour, and further, the device was found highly portable and cost-
effective. Even though Senapati et al. (2009) promoted the simplicity of the device,
the sampling approach contains few fabrication steps, filtration and it requires chem-
ical applications, which inevitably increase the risk of error. Additionally, further re-
search on this method needs to be done in relation to ballast water sampling to estab-
lish the operability of the approach.
Lab-on-chip devices utilizing resistive pulse sensors also require microchip fabrica-
tion, use of primers and loading of the sample (Figure 8) (Song et al., 2012, Wang et
al., 2013). The method utilizes pre-established correlation curve to reduce the sam-
pling time down to 1 – 2 minutes (Song et al., 2012). Song et al. (2012) discovered
that the methodology was sufficiently sensitive to detect the difference in phyto-
plankton concentrations after electrolysis treatment. The device can act as a suitable
replacement for imaging flow cytometers, which tend to be more expensive, time-
consuming and less portable (Wang et al., 2013). Furthermore, separate chips can be
designed to detect organisms in the preferred size categories (>50 µm, 10 µm and
<50 µm or <10 µm) (Song et al., 2012). First and Drake (2013) suggested that lab-
on-chip devices could be potentially installed into piping systems in ships to provide
real-time data from the ballast water.
Trafi Publications 10-2017
25
Figure 8. Illustration of the fluorescence detection system (a), microfluidic chip (b) and detection area (c) (Wang et
al., 2013).
Trafi Publications 10-2017
26
3.7.2 Fast repetition rate (FRR) fluorometry
Closely related to PAM fluorometry, FRR fluorometry measures chlorophyll fluo-
rescence (Kolber et al., 1998). In comparison to PAM fluorometry using multiple
turnovers of PS II from excitation flashes, FRR fluorometry exposes the sample to a
high-energy single turnover flash to obtain the maximum fluorescence level
(Kromkamp and Forster, 2003). The main differences between PAM and FRR fluo-
rometric methods derive from the length of the applied flash pulses. Multiple turno-
ver technique utilizes longer flash times, which can lead to overestimation of chloro-
phyll due to improved detection for background fluorescence (Figure 9) (Kromkamp
and Forster, 2003). Therefore FRR fluorometers are not only more sensitive, but also
more suitable for phytoplankton measurements. Additionally, the method enables
the detection of cyanobacteria by being able to determine smaller functional cross
sections of PS II (Sugget et al., 2001, Sugget et al., 2006).
Figure 9. Multiple turnover technology (blank dots) tends to overestimate the PS II efficiency and electron transport
rate in samples in comparison to single turnover technology (black dots) (Kromkamp and Forster, 2003).
Chelsea Technologies Group has developed a portable and quick FastBallast sam-
pling device that utilizes applied FRR fluorometry for onboard analyses, which has
similar benefits as PAM fluorometers (CTG, 2016). The FastBallast technology
aims to provide a detailed analysis from the concentration of organisms in the <50
µm and 10 µm category, either via discrete sample analysis, flow-through analysis
or through integrated system in real-time. Additionally, FastBallast technology ena-
bles the sampling of larger water volumes and can make a distinction between dif-
ferent cell sizes, as well as background fluorescence (CTG, 2016). This technology
Trafi Publications 10-2017
27
has huge potential for compliance monitoring in the future, as it claims to combine
accuracy of a detailed analysis and simplicity of an indicative analysis method.
However, FastBallast technology and FRR fluorometry have not been studied
widely in compliance-related ballast water studies to this date. Therefore, the relia-
bility of this method requires further confirmation from laboratory and onboard tri-
als.
3.7.3 Serial dilution culture-most probable number (SDC-MPN) method
The SDC-MPN method has been applied to phytoplankton enumeration since the
1950s (Knight-Jones, 1951). Throndsens’ (1978) SDC-MPN method is based on di-
luting samples into a series of subcultures. The concentration of viable cells can be
calculated from the amount of viable cells taking into account the dilution factor
(Figure 10). The sampling steps include the dilution of the sample, incubation period
of 14 days and fluorometer analysis for the detection of the concentration (Casas-
Monroy et al., 2016). Due to the extensive incubation time, this method was not seen
as reasonable of being assessed further.
Figure 10. Approach for the SDC-MPN method (Cullen and MacIntyre, 2016).
Trafi Publications 10-2017
28
3.7.4 Laser transmission spectroscopy (LTS)
In contrast to other light based nanoparticle detection methods measuring light dif-
fraction and scattering, LTS technology detects wavelength-dependent light trans-
mittance through nanoparticles within a sample (Bohren and Huffman, 2000) sum-
marized in Li et al. (2011). The method includes recording of light transmission
through a suspension fluid sample containing nanoparticles, measurement for wave-
length-dependent light transmission and data analysis and inversion by a computer
(Figure 11) (Li et al., 2011). Detailed sampling steps are represented in Li et al.
(2011), but similarly to microfluidic lab-on-chip devices, LTS utilizes nanobeads
with species-specific tags to detect preferred DNA sequences. Therefore, utilization
of LTS technology can be only targeted to individual species, not towards entire or-
ganism categories.
Figure 11. LTS approach for DNA detection (Li et al., 2011).
Several studies, such as Li et al. (2011), Li et al. (2010), Egan et al. (2013) and Egan
et al. (2015) have tested the LTS technology on invasive species detection from bal-
last water. In general, sampling duration for the DNA detection techniques used in
these studies varied widely between 1 min and several hours. The methodology was
proven sensitive and effective in the detection of the targeted DNA, making clear
distinction even to closely related species (Li et al., 2011, Egan et al., 2013, Egan et
al., 2015). The devices include also portable, cost-effective and automated versions
(Egan et al., 2013), and recent development models aim to obtain accurate results in
10 seconds to confirm the compliance status for ships (Egan et al., 2015).
3.7.5 Analysis techniques for larger zooplankton
Bradie (2016) have represented optical zooplankton analysis (OZA) method to de-
tect the concentration of zooplankton larger than 200 µm in minimum dimension
within a sample. The sampling approach is based on swimming capability of these
organisms, as the device captures successive images from the sample and analyses
the results using MATLAB software (Bradie, 2016). Sample is diluted into a bottle
and placed into the OZA device. Results from the first sample, as well as replicates
can be obtained in approximately 15 – 20 minutes, since the sample must be held
still for 15 minutes in the device for the debris to settle down. Bradie (2016) re-
ported that even though the OZA method is portable and relatively fast, its develop-
ment is not yet ready to estimate the acquisition and running costs of the device, and
it highly underestimated organism concentrations in comparison to microscope
counts.
Trafi Publications 10-2017
29
Alternatively, Gollasch (2006) has provided another sampling device to detect zoo-
plankton larger than 50 µm in minimum dimension. The device includes a flow me-
ter, sampling bag and cod-end filter for sample collection (Gollasch, 2006). After the
collection, samples can be analysed with petri dishes and a stereomicroscope. The
device is designed to sample relatively large quantities of water within a short time
range. Additionally, the device is portable and relatively easy to use. However, as
mentioned before, microscopic analysis on zooplankton motility is not the most ap-
propriate method to detect compliance as organism motility cannot be considered as
a sufficient indicator for viability (Bradie, 2016).
3.8 Results overview
In short, feasibility of a sampling method or device for indicative analysis refers to
accuracy and representativeness, sampling duration, ease of use, portability, and
cost-efficiency (David and Gollasch, 2015). Furthermore, the most appropriate sam-
pling methods would include minimal processing and deliver real-time data for mon-
itoring, potentially even directly from the piping system of a vessel (First and Drake,
2013). If real-time analysis methods cannot be achieved, it is essential that the pre-
ferred methods require minimal amount of chemicals and reagents used, as these
methods require processing in a laboratory environment.
The definition for indicative analysis sampling methods in this study includes meth-
ods with maximum analysis time of 2 hours, as this was required by the Finnish
Transport Safety Agency. These methods and their feasibilities are evaluated in Ta-
ble 7. Table 7 was designed to list the methods that should be discussed further in
the discussion. Instead of evaluating the sampling accuracy and representativeness
here, it was done for the most reliable methods in the discussion section. All meth-
ods described in this study are represented in Appendix C (Table 8).
Trafi Publications 10-2017
30
Table 7. Feasibilities of the assessed indicative sample analysis methods. Green colour represents the best possible suitability
for the assessed factor, whereas orange colour refers to the second best suitability and red colour indicates the poorest suita-
bility. Suitability factors are comprehensively described in the discussion. Only the six best methods were ranked into this table
and rest of the methods were considered unsuitable indicative analysis methods at the time of the study. Most information in
this table has been gathered from the literature and some evaluation has been applied on the ‘skills required’ factor. Device
manufacturers include the device manufacturers (e.g. luminometer, fluorometer or epifluorescent microscope), but not neces-
sarily manufacturers for all reagents/chemicals. Method ranking is based on our evaluation about the method suitability, num-
ber 1 indicating the most suitable method and number 6 the least suitable (Gollasch, 2006, Braissant et al., 2010, Reavie et
al., 2010, Welschmeyer and Maurer, 2011, Li et al., 2011, Gollasch et al., 2012, Song et al., 2012, First and Drake, 2013,
Stehouwer et al., 2013, Bakalar, 2014, Nakata et al., 2014, David and Gollasch, 2015, Gollasch et al., 2015, van Slooten et al.,
2015, Bradie, 2016, CTG, 2016).
Method Duration Method reli-
ability
Chemi-
cals re-
quired
Skills
required Portable Cost
Re-
quires
labora-
tory
analysis
Manufacturers Rank
ATP 15 min – 1
hour Highly relia-
ble Yes Some skills
required Yes <10,000
$ No
Aqua-tools
(SGS, Luminul-
tra),
Welschmeyer,
Hygiena
2
FDA 30 min – 2
hours Moderately
reliable Yes Some skills
required Yes 450 –
15,000
$ No
Satake, Turner
Designs,
Horiba, Hach,
Olympus, SPI
Supplies, Sev-
eral manufactur-
ers for counting
devices
4
PAM Few minutes Highly relia-
ble No Minimal Yes 4,000
15,000
$ No
BBE Moldaenke
GmbH, Turner
Designs, Heinz
Walz GmbH,
Hach
1
Flow cy-
tomet-ry Few minutes –
few days Moderately
reliable No Some skills
required No
18,000
$ –
200,000
No
BD Biosciences,
Beckman Coul-
ter, Inc, FluidIm-
aging, Bio-Rad,
Cytobuoy
-
Bacterial
fluoro-me-
ters 10 – 20 min Moderately
reliable Yes Some skills
required Yes 2,000
10,000
No
IDEXX, Bactest,
Triton Marine
Science & Con-
sult, Turner de-
signs, Vista En-
terprises, Inc,
DeltaTrak, Inc
6
Microca-
lorimetry Few hours Minimally re-
liable No Some skills
required Yes N/A N/A N/A -
Colori-
metry 15 – 30 min Moderately
reliable Yes Some skills
required Yes
32,000
151,000
$
No Vitek, New Hori-
zons Diagnos-
tics, Coleparmer -
Lab-on-
chip 1 min – 2
hours Moderately
reliable Yes Some skills
required Yes N/A No Chips con-
structed by vari-
ous pieces 5
FRR fluo-
rometry 2 – 10 minutes Moderately
reliable No Minimal Yes N/A No Chelsea Tech-
nologies Group,
Ltd 3
LTS Few min –
Few hours Minimally re-
liable Yes Some skills
required Yes N/A No Devices include
several items -
OZA 10 – 30 min Minimally re-
liable No Some skills
required Yes N/A No N/A -
Zoo-plank-
ton sam-
pling bag 30 min Minimally re-
liable No Some skills
required Yes N/A No Hydrobios -
Trafi Publications 10-2017
31
4. Discussion
As a PSC officer should be able to conduct the sampling with some training, but
without an academic education for biology or chemistry (David and Gollasch, 2015),
an indicative method should require minimal skills, and the results should be ob-
tained after just placing a sample and pushing a button (Table 7). Reliability of a
method refers to not only representativeness and accuracy of a sample, but also how
widely has the method been tested and evaluated in the previous studies. These fac-
tors ultimately affected our ranking system in Table 7. For example, even though
Adenosine TriPhosphate (ATP) method requires an extraction step, it was consid-
ered more reliable than FRR method, because ATP has been studied widely in bal-
last water related trials. The ranking system was based on personal evaluation of the
methods presented in the existing literature, and it should be highlighted that other
authorities or experts could have constructed the ranking system differently.
In addition, Finnish Transport Safety Agency noted that UV-treatment is clearly the
most recommended method between the Baltic Sea area operators and nearly all
Ballast Water treatment systems (BWTS) include also a filtration down to 40 µm or
even 20 µm (Stehouwer et al., 2015), although filtration alone is probably an insuffi-
cient method to treat ballast water. First and Drake (2014) evaluated that certain in-
dicative ballast water analysis methods can be treatment-specific, potentially having
variability in the detection of BWTS efficiency between different treatment meth-
ods. This attribute of UV-treatment has been taken into consideration in the discus-
sion section, even though some methods have been able to overcome issues related
to compliance monitoring after UV-treatment. Filtration down to 40 µm in turn, pro-
motes the testing of smaller organisms than 40 µm since larger organisms are un-
likely to pass the filtration, unless the Port State Control (PSC) authority prefers to
test the efficiency of filters.
Several methods in Table 7, such as microcalorimetry, colorimetry, lab-on-chip de-
vices, Fast Repetition Rate (FRR) fluorometry, Laser Transmission Spectroscopy
(LTS), Optical Zooplankton Analysis (OZA) and zooplankton sampling bag suffer
from inadequate amount of reliable studies, or alternatively, have not been applied to
ballast water compliance sampling comprehensively.
The main purpose of this section is to determine the most appropriate methods for
indicative analysis. Once the most appropriate method or combination of methods
has been identified, the following section will include also recommendations on the
device manufacturers and device-specific notifications.
4.1 General issues and uncertainties regarding the sampling
methods
The previous section revealed that the evaluated indicative analysis methods include
all some sort of limitations or uncertainties that can potentially have a negative ef-
fect on the accuracy and representativeness of the obtained results. Some of these
limitations have been overturned by better technology, but some methods may con-
tain relatively high uncertainties within the sampling approach. If these uncertainties
or limitations cannot be overturned, the method can be considered as unsuitable for
compliance testing, and there is no need to evaluate the method any further.
Trafi Publications 10-2017
32
4.1.1 Minimally reliable methods
As described in Table 7, the OZA, zooplankton sampling bag, microcalorimetry and
LTS methods were all considered as minimally reliable for compliance testing. The
OZA method is quite early in the development stages (Bradie, 2016) and has not
been used widely for compliance testing. Therefore it cannot be considered as a reli-
able indicative analysis method in the present study. In order to obtain results from
zooplankton sampling bag method, as well as methods assessing smaller zooplank-
ton, microscopic analysis on zooplankton motility is required (Gollasch, 2006) and it
has been noted repeatedly in the present study that this requires laboratory environ-
ment and organism motility is not an appropriate measure for viability (Bradie,
2016).
The main issue related to microcalorimetry is that sampling duration is depending on
heat production from metabolic activity of the targeted organism size category (First
and Drake, 2013). For the detection of bacterial concentration, the process of heat
production requires several hours for smaller organisms (Wadsö, 2002). The analy-
sis time, also known as the size-dependent heat production for larger organisms
takes only an hour, but sampling protocol for these organisms requires extensive
sample preparation (up to several days) (Johnson et al., 2009). Therefore these meth-
ods cannot be recommended for indicative compliance testing.
LTS technology has been studied quite extensively in association with ballast water
sampling by Li et al. (2010), Li et al. (2011), Egan et al. (2013) and Egan et al.
(2015). The method is able to detect the presence or absence of the targeted species
quickly and effectively (Li et al., 2011, Egan et al. 2015). Unfortunately the ap-
proach shares the same issue with other species-specific sampling methods by being
able to only concentrate on single species. The result represents therefore only the
absence or presence of the targeted species and fails to indicate the compliance for
all other organisms even in one size category. This disadvantage is difficult to over-
turn and LTS approach alone can be considered as insufficient analysis method for
compliance determination.
4.1.2 Moderately reliable methods
According to Table 7, moderately reliable methods in this study included Fluores-
cein diacetate (FDA), flow cytometry, bacterial fluorometers, colorimetry and lab-
on-chip devices. Aforementioned statement “any type of staining method has the is-
sue of not being able to differentiate between living and viable cells (Reavie et al.,
2010)” has been found particularly valid for variable FDA methods in the present
study. Several studies, such as Tobiesen et al. (2011), Steinberg et al. (2011), Adams
et al. (2014) and MacIntyre and Cullen (2016) all reported that FDA, carboxyfluo-
rescein diacetate (CFDA) and 5-Chloromethylfluorescein diacetate (CMFDA) stains
tended to overestimate the viable organism concentrations by staining dead cells.
Counting the cell concentrations using an epifluorescent microscope exposes the
staining methods for live and dead cells to errors, originating from the prolonged
counting processes (First and Drake, 2013). The portable FDA bulk analysis con-
ducted by Welschmeyer and Maurer (2011) in turn, suffered from extensive sam-
pling duration (3 hours).
The only FDA method that could be potentially recommended for indicative analysis
sampling is the FDA pulse counter represented by Nakata et al. (2014). The FDA
Trafi Publications 10-2017
33
pulse counting method is very simple and Nakata et al. (2014) obtained great results
using the device. However, it is difficult to rely on a single study while evaluating
the suitability of this method. Bradie (2016) was the other researcher who applied
the FDA pulse counting method in her study. She concluded that the device underes-
timated the cell counts in comparison to microscopic inspection. The device has po-
tential of becoming a reliable indicative analysis tool, but the lack of associated sci-
entific studies on its efficiency and reliability does not promote the recommendation
of such device. Therefore FDA methods cannot be recommended for further evalua-
tion on reliable or suitable indicative analysis methods yet.
Interestingly, even though FRR fluorometry has relatively similar sampling ap-
proach with Pulse Amplitude-Modulation (PAM) fluorometry including mutual fea-
sibility benefits, such as portability, quickness and simplicity, the methodology has
not been assessed or used widely in ballast water trials or studies. Studies by Sugget
et al. (2001), Kromkamp and Forster (2003) and Sugget et al. (2006) have reported
that FRR approach has sensitivity and accuracy benefits over PAM technology re-
garding marine phytoplankton samples. Furthermore, the FastBallast technology uti-
lizing applied FRR fluorometry can be a successful compliance monitoring tool in
the future if the reported attributes of the device can be scientifically proven. The
greatest concern with FRR fluorometry at present relates to lack of ballast water re-
lated studies. The technology has been developed almost 20 years ago (Kolber et al.,
1998) and if it is indeed superior to PAM fluorometry, which is by far the most rec-
ommended indicative analysis method out there, why has the technology not been
developed further in D2 compliance monitoring? However, FRR fluorometry cannot
be ignored from promising indicative analysis methods, as it has potential character-
istics to be an efficient sampling methodology. Further evidence on its applicability
on ballast water compliance monitoring is required to confirm the comprehensive
suitability of the method.
Flow cytometry method has the benefit of being able to provide accurate results on
organism concentrations (Bakalar, 2014) and therefore it is recommended as a pri-
mary method for detailed analysis (David and Gollasch, 2015). Even though quick
flow cytometers also appear to be available (Stehouwer et al., 2013), it remains un-
clear how accurately these devices are able to differentiate viable cells from non-via-
ble (Olsen et al., 2015, Bradie, 2016). Stehouwer et al. (2013) and Bakalar (2014)
already highlighted relatively expensive prices of flow cytometers, but more im-
portant aspect appears to be portability of these devices. After a short exploration on
technical specifications of the devices used in Joachimsthal et al. (2003), Peperzak
and Brussaard (2011) Stehouwer et al. (2013), Bakalar (2014) and Olsen et al.
(2015), it became evident that these devices weigh roughly between 20 and 100+
kilograms, indicating that they are not easily portable. These factors altogether sug-
gest that flow cytometers are not the most appropriate devices for indicative analysis
sampling.
Method for detection of bacterial enzymes within 2 hours suffers also from rela-
tively limited studies associated with ballast water sampling, as mainly Gollasch et
al. (2012) and Bradie (2016) have reported about the operability of such devices.
Relatively simple, portable and fast devices are indeed available (Gollasch et al.,
2012) that are able to provide indicative results. As PAM and ATP methods are una-
ble to directly indicate the presence or absence of Regulation D2 bacteria, handheld
fluorometers can be recommended if the presence of bacteria explicitly needs to be
assessed. However, the literature associated with bacterial fluorometers as indicative
Trafi Publications 10-2017
34
ballast water analysis devices was found relatively limited and the operability of
these devices is suggested to be tested further. In addition, this methodology cannot
be recommended as the only indicative sampling device, since it is designed to only
assess the targeted bacterial enzymes and therefore fails to indicate the presence of
other organisms within the ballast water.
Detection for bacteria via colorimetry method has been discussed by Gollasch et al.
(2012) and Bakalar (2014). Both studies stated that a simple colorimetric assessment
for Regulation D2 bacteria can be done within 15 minutes, although Gollasch et al.
(2012) noted uncertainties regarding to detection limit for bacteria and operability in
varying salinities. After focusing on the details of the devices assessed in these stud-
ies, it was revealed that the device evaluated in Bakalar (2014) cannot be considered
as portable, weighing approximately 75 kilograms (Vitek, 2016). Similarly, devices
evaluated in Gollasch et al. (2012) have been mainly designed to assess stool sam-
ples (New Horizons Diagnostics, 2011), without distinct relation to testing of ballast
water. Due to the lack of ballast water related studies and applications using color-
imetry, the methodology cannot be recommended as a suitable option for indicative
analysis.
Similarly to bacterial enzyme detection, lab-on-chip devices are mainly able to ana-
lyse only the presence of the DNA-targeted species, even though they can also de-
tect chlorophyll fluorescence (Song et al., 2012). However, PAM fluorometry is sig-
nificantly faster and easier to use in comparison to lab-on-chip methodology for
chlorophyll fluorescence measurements. The actual measurement is simple with lab-
on-chip devices, but the chip fabrication increases the complexity of the method.
The method suffers from lack of studies executed outside of laboratory, but has pro-
vided promising results in studies by Song et al. (2012), Wang et al. (2013) and
Song et al. (2014). In short, it would be difficult to recommend lab-on-chip devices
over PAM and bacterial fluorometers, but if automated chips can be constructed and
installed into ships’ piping system to deliver real-time data, this technology should
be definitely developed and supported further (First and Drake, 2013).
Approaches targeting individual species cannot be recommended alone, as there is
no universal species always present in ballast water that would indicate the compli-
ance status. Same principle applies to detection of D2 bacteria, which can be tested
in an indicative manner by using bacterial fluorometers. However, if an extremely
harmful species, e.g. a target species is recognized which entry into the country
would be aimed to prevent and the sampling is preferred to indicate the presence of
this species alone, aforementioned methods targeting individual species can be ap-
plied. This kind of case is however not related to the compliance of the vessel and
thus does not indicate the ballast water treatment success.
The current target species list for the Baltic Sea area is presented in Appendix D
(Table 9) (HELCOM, 2015). Majority of the species in this list belong to the cate-
gory greater or equal to 50 µm in minimum dimension, and therefore sampling for
these species would target most of the unwanted species. Anyhow, it is essential to
understand that if the sampling is targeted to target species, or any other group of
species, the process does not apply for compliance monitoring purposes. As pre-
sented in the introduction, organisms smaller than 50 µm and greater or equal to 10
µm are the most reliable indicator group for compliance monitoring (David and Gol-
lasch, 2015, Gollasch et al., 2015), thus leaving most of the species on the target list
outside of the sampling.
Trafi Publications 10-2017
35
4.1.3 Highly reliable methods
ATP and PAM methods were considered as highly reliable indicative analysis meth-
ods in the present study. The essential benefit of the ATP method is the ability to
measure and detect all organism categories listed in regulation D2 of the Ballast Wa-
ter Management Convention (BWM Convention) (van Slooten et al., 2015, Bradie,
2016). The main issues related to ATP method refer to sampling accuracy, as the
presence of inorganic constituents within a sample and potential effects of ultraviolet
(UV)-treatment can influence the obtained results (Villaverde et al., 1986, Sudharan
and Reddy, 2000). However, more recent studies, such as Penru et al. (2012), Wright
et al. (2015) and Welschmeyer and Kuo (2016) have been able to overcome these
sampling accuracy related issues, suggesting that current ATP methods can success-
fully detect the presence and absence of organisms from ballast water although UV-
treatment would have been applied. Therefore ATP method is evaluated also further
in the following sections.
PAM fluorometry has been widely recommended as potentially the best indicative
analysis sampling method (David and Gollasch, 2015), in terms of cost-efficiency,
quickness, portability and technical simplicity (First and Drake, 2013, Bakalar,
2014, van Slooten et al., 2015, Bradie, 2016). PAM method can also be vulnerable
to under- or overestimation of organism concentrations (Casas-Monroy et al., 2016)
and has limitations in association with measuring only autotroph concentrations (van
Slooten et al., 2015). Even though PAM measurements are somewhat limited to de-
tect only the presence of phytoplankton, van Slooten et al. (2015) and Bradie (2016)
were able to obtain accurate results with PAM devices. Due to the conclusions of
these studies, PAM method can also be recommended as a suitable indicative analy-
sis method.
Overall, all of the aforementioned methods have their advantages and disadvantages.
These methods have been evaluated according to the preferences presented by the
Finnish Transport Safety Agency, with the purpose of identifying the most feasible
method, but nevertheless, avoiding the feasibility to impact negatively on sampling
accuracy and reliability. Based on the feasibility and suitability of the assessed meth-
ods, PAM and ATP can be evaluated further in this report as the most appropriate
indicative analysis methods.
4.2 Accuracy
Sampling accuracy relates to representativeness and reliability of the evaluated ap-
proach. The benefit of ATP method over PAM fluorometry is the ability to evaluate
the concentration of all organism size categories including autotrophic and hetero-
trophic organisms (van Slooten et al., 2015, Bradie, 2016). PAM fluorometry, in
turn, is only able to detect the concentrations of autotrophs through chlorophyll a
analysis, which can exclude certain organisms even from the measured <50µm and
10µm size class (Gollasch et al., 2015). In theory, no matter how accurately PAM
fluorometry is able to measure and count the presence of viable phytoplankton cells,
the outcome cannot be considered as being fully representative of this organism size
class. In practice however, if the ballast water is dominated by phytoplankton spe-
cies (Bradie, 2016), PAM fluorometry can provide reliable measurements on the
compliance status of the ship.
Trafi Publications 10-2017
36
Even though PAM measurements are somewhat limited to detect only the presence
of phytoplankton, van Slooten et al. (2015) and Bradie (2016) were able to obtain
accurate results with PAM devices. Furthermore, Gollasch et al. (2015) and Casas-
Monroy et al. (2016) were able to occasionally obtain more reliable results using
PAM fluorometry than the ATP method. Most PAM devices require calibration in
order to transfer the fluorescence values into organisms per volume (Bradie, 2016).
This can be relatively problematic, as there is no single cell size in the <50µm and
10µm size class that would provide universal conversion into organism concentra-
tions. Therefore device comparison is important, as devices can present different or-
ganism concentration estimates with the same fluorescence reading.
ATP extraction can be conducted by using a wide range of buffers, acids and sol-
vents (Karl, 1980). The extracting reagent can have a significant impact on the ob-
tained results (Welschmeyer and Kuo, 2016), as the ATP method requires extraction
of ATP from the cells within a sample (First and Drake, 2014). Extraction tech-
niques effect on results, as well as on sampling feasibility, since some extraction
methods require laboratory resources (Welschmeyer and Kuo, 2016). Stronger acid
extraction and various extraction agents are usually preferred on more complex sam-
ples containing various organisms. Welschmeyer and Kuo (2016) studied the effi-
ciencies of 3 different ATP extraction techniques on cultured phytoplankton. De-
tected ATP concentrations varied significantly between the extraction techniques
(Figure 12). However, all of these extraction techniques detected successfully the
decreased ATP concentration in samples after UV-treatment in the study.
Figure 12. Efficiencies of 3 different ATP extraction methods (Welschmeyer and Kuo, 2016).
Trafi Publications 10-2017
37
Interpretation of the results can be somewhat problematic, as for organisms smaller
than 50 µm and greater or equal to 10 µm, Aqua-tools (2016) suggests that 500 ng/l
of cellular ATP is a critical limit for the compliance of the ship, whereas
Welschmeyer and Kuo (2016) express that 26.2 ng/l would be the limit value.
Welschmeyer and Kuo (2016) used the same LuminUltra Photon Master luminome-
ter that is promoted by Aqua-tools (2016), even though this limit value should be in-
dependent of analysis device, as it measures concentration of cellular ATP for the
same size class.
Therefore, more research should be targeted on critical ATP concentration limits de-
tecting compliance status. It is understandable that compliance limits of ATP have
some variation, since the ATP extraction efficiency also varies between the used ex-
traction techniques and reagents. Wright et al. (2015), Bradie (2016) and
Welschmeyer and Kuo (2016) studied ATP concentrations before and after ballast
water treatment. In these studies, the ATP concentrations for the <50 µm and 10
µm category in untreated ballast water varied between 60 and 914 ATP ng/l,
whereas after treatment the ATP concentrations varied between 0.07 and 20 ATP
ng/l. This finding suggests that when BWTS is working, the ATP concentration in
ng/l for this size group should be clearly less than 25. On the contrary, if the same
concentrations are above 25, the operability of BWTS can be questioned, as
Welschmeyer and Kuo (2016) tested the ATP concentrations from various locations
after 4 different treatments using different extraction agents and never recorded a
value higher than 11.61 ATP ng/l after treatment.
Even though the ATP method enables the sampling of all organism size categories,
it is generally used to estimate the organism concentration in the <50 µm and 10
µm category, since sampling for larger organisms requires significantly larger sam-
ple volumes (Miller et al., 2011) and Regulation D2 bacteria are rarely present even
in untreated water (Welschmeyer and Kuo, 2016). However, Bradie (2016) used
ATP method to detect organisms in the 50 µm category. Her study revealed that
the ATP method showed relative correlation to microscopy counts (r= 0.552), and
was able to detect significant decrease in organism concentration (50 µm) after bal-
last water treatment.
4.3 Manufacturers
As mentioned in the previous section, due to the differences in calibration, operabil-
ity and device specifications, further research on PAM and ATP devices between
manufacturers is needed. The selection between ATP equipment manufacturers was
not found widely variable in the present study. The main devices used in ballast wa-
ter related studies were aqua-tools 2G ATP kit, Welschmeyer ATP and Sys-
temSURE ATP instrument by Hygiena. Aqua-tools developed their device in collab-
oration with SGS and LuminUltra (Aqua-tools, 2016). The SystemSURE equipment
was primarily designed to detect presence of microbes in municipal waters (Hy-
giena, 2016). However, First and Drake (2014) used the Hygiena device in a ballast
water study and they detected a significant decrease in ATP after chlorine dioxide
treatment but not after UV treatment, possibly due to aforementioned effects of UV
treatment potentially increasing the amount of cellular ATP in bacteria (Villaverde
et al., 1986). Unfortunately, no further details were found on the Welschmeyer ATP
method used in Bradie (2016), which prevented the evaluation of this option. There
Trafi Publications 10-2017
38
were also various other luminometer and extracting agent manufacturers, but their
equipment was not necessarily specified to ballast water compliance monitoring.
Gollasch et al. (2015) and Bradie (2016) compared 4 PAM devices in their studies,
which can be identified as the main PAM devices designed for ballast water compli-
ance monitoring. Similarly to existing ATP luminometers, there were also other
PAM fluorometer manufacturers, but their devices were mainly specified to chloro-
phyll recordings from leaves. Manufacturers for the devices studied by Gollasch et
al. (2015) and Bradie (2016) were BBE Moldaenke GmbH, Turner Designs, Heinz
Walz GmbH and Hach Corporation. Gollasch et al. (2015) reported that the PAM
device manufactured by Hach Corporation provided the most consistent results be-
tween replicates and was able to match the risk levels correctly with the altered, mi-
croscopically inspected phytoplankton counts.
Bradie (2016) obtained relatively different results. All 4 PAM devices correlated
highly with microscope counts in the <50µm and 10µm size class with minimal
differences (r= 0.82 – 0.87) including significantly high correlations between the de-
vices (r= 0.75 – 0.96). Moreover, study by Bradie (2016) revealed that the sensitiv-
ity varied between the devices with Hach BW680 being the least sensitive device.
Hach BW680 was not able to detect fluorescence signal in any replicates within 17
trials, where organisms were found present by microscopic inspection, whereas
other devices were able to record signals from these lower chlorophyll concentra-
tions. However, differences between the PAM devices in this size class were rela-
tively minimal and correlations with the devices and microscope counts varied re-
markably in other organism categories. In terms of feasibility, specifications of these
ATP and PAM devices for comparison are presented in Table 10. Certain PAM de-
vices, such as Ballast-Check 2 by Turner Designs and Hach BW680 are factory cali-
brated and ready for use immediately, but despite some differences in sampling pro-
tocols between these devices, none of them was considered difficult to use.
Table 10. Specifications of the ATP and PAM devices compared in the present study. Running costs are presented
in brackets. These costs for Hach PAM, Turner Designs PAM and SystemSURE ATP were not available (Gollasch
et al., 2015 and Bradie, 2016).
Manufacturer Walz
(PAM)
Turner
Designs
(PAM)
BBE (PAM) Hach
(PAM)
Aqua
-
tools
(ATP)
SystemSURE
(ATP)
Weight < 5kg +
computer 0.4 kg 5 kg 0.3 kg N/A 0.3 kg
Analysis time 5 min 1 min 2-3 min 1 min 15 – 50
min 1 min
Price 15,000 €
(5€/sample) 4,500 –
5,000 $
4,300 –
13,800 €
(1€/sample)
<
5,000
$
7700€
(13.33 –
31.83
€/sample)
2,000 – 2,500 €
Filtration re-
quired No No Yes No Yes Yes
Computer
use External No Internal No External Optional
Trafi Publications 10-2017
39
The differences in device feasibility or accuracy are relatively minimal, as can be
seen from Table 9 and the associated studies by Gollasch et al. (2015) and Bradie
(2016). It makes sense that the most consistent results are obtained by the least sen-
sitive device, since it is most likely unable to detect small differences between sam-
ple replicates. All of the devices are reasonably priced, can be easily used and oper-
ate quickly, although Bradie (2016) presented that the Aqua-tools ATP device in-
cludes additional running costs (31.83 € for organisms <50µm and 10µm and 50
µm and 13.33 € for bacteria per sample). PAM fluorometers have variation in the or-
ganism count determination (Bradie, 2016), but all devices have been able to detect
significant differences after ballast water treatment compared to non-treated waters.
As PAM fluorometers are able to detect only the viability of phytoplankton, a com-
bination of PAM and ATP tests should deliver a reliable indication of the compli-
ance. Particularly, as the statement by Bradie (2016) on the presence of phytoplank-
ton being able to indirectly detect the presence of zooplankton explained by lack of
predation can be questioned. If phytoplankton is detected absent in ballast water
samples, it can be simply because the treatment system has worked, or alternatively,
due to predation by zooplankton, or both.
4.4 Study limitations
The outcomes of the present study are primarily limited to existing literature and
consultation with other researchers. The device attributes and methodological steps
conducted in the previous studies can be evaluated to certain extent, but literature
review does not compare directly studies including onboard or lab trials. It is essen-
tial to notify that most of the studies cited in this literature review were also limited
to studying only one or few indicative methods at a time, or alternatively, the sam-
pling devices were tested on ballast water originating from certain areas, where the
water was dominated by certain group of species.
Drawn conclusions from comparing individual studies require special caution and
further research on compliance monitoring is recommended, preferably including
various analysis methods on ballast water originating from various locations. In ad-
dition, more attention should be drawn to investigation of methods that are able to
provide reliable real-time data on organism viability from ballast water tanks or
pipes, which would decrease the burden of PSC officers significantly.
5. Conclusions and recommendations
The main goal of the current study was to present the existing indicative ballast wa-
ter analysis methods for compliance monitoring in Finland. Additionally, the present
study aimed to identify the best sampling methods and devices in terms of reliabil-
ity, accuracy and feasibility. The recommended methods in the present study should
therefore be able to combine the advantages of low price, quick analysis time, porta-
ble analysis devices and reliable results without requiring an academic education for
biology or chemistry (David and Gollasch, 2015).
The outcomes of the present study revealed that indicative ballast water analysis
should be conducted using the combination of PAM fluorometry and ATP method.
Both of these methods are able to detect the effect of ballast water treatment in a
Trafi Publications 10-2017
40
simple, quick, portable and relatively cheap manner. Turbid water, a relatively com-
mon attribute for Baltic Sea waters (Granqvist and Mattila, 2004) should not cause
significant disturbances for the results obtained by PAM and ATP devices (Waite et
al., 2003, Aqua-tools, 2016, Bradie, 2016). Utilization of 2 different methods also
increases the reliability of sampling, when the presence of different sizes and types
of organisms can be detected. The differences between various PAM and ATP de-
vices were not considered significant in the present study. It is primarily up to PSC
authorities, whether they want to invest on device sensitivity, or prefer relatively
cheaper and more portable options.
David and Gollasch (2015) and Gollasch et al. (2015) determined that organisms
<50µm and 10µm are the most reliable indicator group for the compliance moni-
toring and presence of these organisms can be detected by using both, PAM and
ATP method. If PSC authorities prefer focusing the sampling process towards cer-
tain harmful species, methods targeting individual species or group of species can be
used (e.g. LTS), but due to being unable to detect the presence of other species, the
outcomes off these approaches are not representative for the full compliance status
of the vessel. Overall, more research on compliance monitoring is suggested, espe-
cially using several analysis methods on ballast water loaded from various locations
to identify the most efficient and reliable indicative analysis method transparently.
6. Acknowledgements
The present study consisted of searching relevant literature and consulting experts
about the existing indicative analysis methods and devices. We would like to thank
Finnish Transport Safety Agency, Special Adviser Ville-Veikko Intovuori, Head of
Unit Mirja Ikonen and Chief Adviser Dr. Anita Mäkinen for enabling the execution
of the present study. Additionally, we want to express our gratitude to experts
Stephan Gollasch, Christian Moldaenke, Marcel Veldhuis, Sarah Bailey, Tom
Brumett and Stephanie Lavelle on their advises on methodological and technological
details.
Trafi Publications 10-2017
41
7. Bibliobraphy
Adam, G., & Duncan, H. (2001). Development of a sensitive and rapid method for
the measurement of total microbial activity using fluorescein diacetate (FDA) in a
range of soils. Soil Biology and Biochemistry, 33(7), 943-951.
Adams, J., Briski, E., Ram, J. L., & Bailey, S. A. (2014). Evaluating the response of
freshwater organisms to vital staining. Management of Biological Invasions, 5(3),
197-208.
Akram, A. C., Noman, S., Moniri-Javid, R., Gizicki, J. P., Reed, E. A., Singh, S. B.,
Basu, A. S., Banno, F., Fujimoto, M. & Ram, J. L. (2015). Development of an auto-
mated ballast water treatment verification system utilizing fluorescein diacetate hy-
drolysis as a measure of treatment efficacy. Water research, 70, 404-413.
Albert, R. J., Lishman, J. M., & Saxena, J. R. (2013). Ballast water regulations and
the move toward concentration-based numeric discharge limits. Ecological Applica-
tions, 23(2), 289-300.
Aqua-tools. (2016). Ballast water monitoring kit. Available from: http://www.aqua-
tools.com/en/ballast-water/
Azam, F., & Hodson, R. E. (1977). Dissolved ATP in the sea and its utilisation by
marine bacteria. Nature, 267, 696-698.
Bakalar, G. (2014). Review of interdisciplinary devices for detecting the quality of
ship ballast water. SpringerPlus, 3(1), 1.
Bakalar, G. (2011). Efforts to develop a ballast water monitoring device. In Second
IMO Globallast R&D Forum on Compliance Monitoring and Enforcement.
Berges, J. A., & Falkowski, P. G. (1998). Physiological stress and cell death in ma-
rine phytoplankton: induction of proteases in response to nitrogen or light limitation.
Limnology and Oceanography, 43(1), 129-135.
Binet, M. T., & Stauber, J. L. (2006). Rapid flow cytometric method for the assess-
ment of toxic dinoflagellate cyst viability. Marine environmental research, 62(4),
247-260.
Bohren, C. F., & Huffman, D. R. (2000). Absorption and scattering of light by small
particles. John Wiley & Sons.
Bradie, J. (2016). METEOR Voyage M116/2: Report on performance of ballast wa-
ter collection and analysis devices. Prepared for BSH (German Federal Maritime
and Hydrographic Agency): 130 pages.
Braissant, O., Wirz, D., Göpfert, B., & Daniels, A. U. (2010). Use of isothermal mi-
crocalorimetry to monitor microbial activities. FEMS microbiology letters, 303(1),
1-8.
Brussaard, C. P., Marie, D., & Bratbak, G. (2000). Flow cytometric detection of vi-
ruses. Journal of virological methods, 85(1), 175-182.
Budnick, G. E., Howard, R. T., & Mayo, D. R. (1996). Evaluation of Enterolert for
enumeration of enterococci in recreational waters. Applied and environmental mi-
crobiology, 62(10), 3881-3884.
Cangelosi, A. (2011). Final Report of the Land-Based, Freshwater Testing of the
Lye (NaOH) Ballast Water Treatment System.
Carney, K. J., Basurko, O. C., Pazouki, K., Marsham, S., Delany, J. E., Desai, D. V.,
Anil, A. C. & Mesbahi, E. (2013). Difficulties in obtaining representative samples
for compliance with the Ballast Water Management Convention. Marine pollution
bulletin, 68(1), 99-105.
Trafi Publications 10-2017
42
Casas-Monroy, O., Chan, P. S., Linley, R. D., Byllaardt, J. V., Kydd, J., & Bailey, S.
A. (2016). Comparison of three techniques to evaluate the number of viable phyto-
plankton cells in ballast water after ultraviolet irradiation treatment. Journal of Ap-
plied Phycology, 1-10.
Cordell, J. R., Lawrence, D. J., Ferm, N. C., Tear, L. M., Smith, S. S., & Herwig, R.
P. (2009). Factors influencing densities of non-indigenous species in the ballast wa-
ter of ships arriving at ports in Puget Sound, Washington, United States. Aquatic
Conservation: Marine and Freshwater Ecosystems, 19(3), 322-343.
Costa, E. G., Lopes, R. M., & Singer, J. M. (2015). Implications of heterogeneous
distributions of organisms on ballast water sampling. Marine pollution bulletin,
91(1), 280-287.
CTG, (2016). FastBallast Portable Ballast Water Compliance Monitor. Available
from: http://www.chelsea.co.uk/allproduct/green-shipping/fastballast-portable-bal-
last-water-compliance-monitor#specification
Cullen, J. J., & MacIntyre, H. L. (2016). On the use of the serial dilution culture
method to enumerate viable phytoplankton in natural communities of plankton sub-
jected to ballast water treatment. Journal of applied phycology, 28(1), 279-298.
David, M. (2013). Ballast water sampling for compliance monitoring – Ratification
of the Ballast Water Management Convention. Final report of research study for
WWF International. 66 pp.
David, M., & Gollasch, S. (Eds.). (2015). Global Maritime Transport and Ballast
Water Management: Issues and Solutions (Vol. 8). Springer.
David, M., Gollasch, S., & Leppäkoski, E. (2013). Risk assessment for exemptions
from ballast water management–The Baltic Sea case study. Marine pollution bulle-
tin, 75(1), 205-217.
De Lafontaine, Y., Despatie, S. P., Veilleux, E., & Wiley, C. (2009). Onboard ship
evaluation of the effectiveness and the potential environmental effects of PERA-
CLEAN® Ocean for ballast water treatment in very cold conditions. Environmental
toxicology, 24(1), 49-65.
Dorsey, J., Yentsch, C. M., Mayo, S., & McKenna, C. (1989). Rapid analytical tech-
nique for the assessment of cell metabolic activity in marine microalgae. Cytometry,
10(5), 622-628.
Egan, S. P., Barnes, M. A., Hwang, C. T., Mahon, A. R., Feder, J. L., Ruggiero, S.
T., Tanner, C. E., & Lodge, D. M. (2013). Rapid invasive species detection by com-
bining environmental DNA with light transmission spectroscopy. Conservation Let-
ters, 6(6), 402-409.
Egan, S. P., Grey, E., Olds, B., Feder, J. L., Ruggiero, S. T., Tanner, C. E., & Lodge,
D. M. (2015). Rapid molecular detection of invasive species in ballast and harbor
water by integrating environmental DNA and light transmission spectroscopy. Envi-
ronmental science & technology, 49(7), 4113-4121.
Engström-Öst, J., & Mattila, J. (2008). Foraging, growth and habitat choice in turbid
water: an experimental study with fish larvae in the Baltic Sea. Marine Ecology Pro-
gress Series, 359, 275-281.
First, M. R., & Drake, L. A. (2014). Life after treatment: detecting living microor-
ganisms following exposure to UV light and chlorine dioxide. Journal of applied
phycology, 26(1), 227-235.
First, M. R., & Drake, L. A. (2013). Approaches for determining the effects of UV
radiation on microorganisms in ballast water. Manag Biol Inv, 4, 87-99.
Trafi Publications 10-2017
43
First, M. R., Robbins-Wamsley, S. H., Riley, S. C., & Drake, L. A. (2016). Towards
minimizing transport of aquatic nuisance species in ballast water: Do organisms in
different size classes respond uniformly to biocidal treatment?. Biological Invasions,
18(3), 647-660.
First, M. R., Robbins-Wamsley, S. H., Riley, S. C., Fisher, J. I., Smith, J. P., &
Drake, L. A. (2014). Examination of additives used to augment “challenge water”
used in verification testing of ballast water management systems: mass yields and
biological impacts. Manag Biol Invasions, 5, 395-405.
Frazier, M., Miller, A. W., Lee, H., & Reusser, D. A. (2013). Counting at low con-
centrations: the statistical challenges of verifying ballast water discharge standards.
Ecological Applications, 23(2), 339-351.
Fykse, E. M., Nilsen, T., Nielsen, A. D., Tryland, I., Delacroix, S., & Blatny, J. M.
(2012). Real-time PCR and NASBA for rapid and sensitive detection of Vibrio chol-
erae in ballast water. Marine pollution bulletin, 64(2), 200-206.
Garvey, M., Moriceau, B., & Passow, U. (2007). Applicability of the FDA assay to
determine the viability of marine phytoplankton under different environmental con-
ditions. Marine Ecology Progress Series, 352, 17-26.
Gasol, J. M., & Del Giorgio, P. A. (2000). Using flow cytometry for counting natu-
ral planktonic bacteria and understanding the structure of planktonic bacterial com-
munities. Scientia Marina, 64(2), 197-224.
Genty, B., Briantais, J. M., & Baker, N. R. (1989). The relationship between the
quantum yield of photosynthetic electron transport and quenching of chlorophyll flu-
orescence. Biochimica et Biophysica Acta (BBA)-General Subjects, 990(1), 87-92.
Gollasch, S. (2012). Quality Management Plan and Quality Assurance Project Plan
for Shipboard Tests of the SiCURE Ballast Water Treatment System. GOconsult,
Hamburg. Version VIII.
Gollasch, S. (2006). A new ballast water sampling device for sampling organisms
above 50 micron. Aquatic Invasions, 1(1), 46-50.
Gollasch, S. & David, M. (2015). Port State Control & Ballast Water Sampling. 4.
VDR –Ballastwasser – Symposium. Hamburg & Leer.
Gollasch, S. & David, M. (2010). Testing sample representativeness of a ballast wa-
ter discharge and developing methods for indicative analysis. Report No. 4. Re-
search Study. European Maritime Safety Agency, EMSA. EMSA, Lisboa, Portugal.
Gollasch, S., David, M., Francé, J., & Mozeti, P. (2015). Quantifying indicatively
living phytoplankton cells in ballast water samples—recommendations for Port State
Control. Marine pollution bulletin, 101(2), 768-775.
Gollasch, S., Stehouwer, P. P., & David, M. (2012). BWO Technical outline and re-
quirements for organism detection systems for establishing compliance enforcement.
Summary of Final Report, Interreg IVB Project, 5.
Gorokhova, E., Mattsson, L., & Sundström, A. M. (2012). A comparison of TO-
PRO-1 iodide and 5-CFDA-AM staining methods for assessing viability of plank-
tonic algae with epifluorescence microscopy. Journal of microbiological methods,
89(3), 216-221.
Granqvist, M., & Mattila, J. (2004). The effects of turbidity and light intensity on the
consumption of mysids by juvenile perch (Perca fluviatilis L.). In Biology of the Bal-
tic Sea (pp. 93-101). Springer Netherlands.
Haugland, R. A., Siefring, S. C., Wymer, L. J., Brenner, K. P., & Dufour, A. P.
(2005). Comparison of Enterococcus measurements in freshwater at two recreational
Trafi Publications 10-2017
44
beaches by quantitative polymerase chain reaction and membrane filter culture anal-
ysis. Water research, 39(4), 559-568.
HELCOM (2015). List of Target Species currently in use in the HELCOM area.
HELCOM, Klaipeda.
Heraud, P., & Beardall, J. (2000). Changes in chlorophyll fluorescence during expo-
sure of Dunaliella tertiolecta to UV radiation indicate a dynamic interaction between
damage and repair processes. Photosynthesis Research, 63(2), 123-134.
Hodson, R. E., Holm-Hansen, O., & Azam, F. (1976). Improved methodology for
ATP determination in marine environments. Marine Biology, 34(2), 143-149.
Hygiena (2016). SystemSURE Plus. Hygiena, LLC. Available from: http://www.hy-
giena.com/systemsure-healthcare.html
Hwang, C. Y., Jung, S., Hwang, Y. S., & Cho, B. C. (2010). Lethal effects of pulsed
high-voltage discharge on marine plankton and Escherichia coli. Water, Air, & Soil
Pollution, 213(1-4), 161-169.
IMO (2016). Status of Conventions. Available from:
http://www.imo.org/en/About/Conventions/StatusOfConventions/Pages/De-
fault.aspx.
IMO (2015). Guidance on ballast water sampling and analysis for trial use in accord-
ance with the BWM Convention and Guidelines (G2). BWM.2/Circ.42/Rev.1. Lon-
don.
IMO (2009). Ballast Water Management Convention. 2009 edition. London, Inter-
national Maritime Organization.
Joachimsthal, E. L., Ivanov, V., Tay, J. H., & Tay, S. T. L. (2003). Flow cytometry
and conventional enumeration of microorganisms in ships’ ballast water and marine
samples. Marine pollution bulletin, 46(3), 308-313.
Johnson, M. D., Völker, J., Moeller, H. V., Laws, E., Breslauer, K. J., & Falkowski,
P. G. (2009). Universal constant for heat production in protists. Proceedings of the
National Academy of Sciences, 106(16), 6696-6699.
Jorgensen, C., Gustavson, K., Hansen, J.B. & Hies, T. (2010). Development of guid-
ance on how to analyze a ballast water sample. Tender N° EMSA/NEG/10/2010.
Karl, D. M. (1993). Total Microbial Biomass Estimation Derived from the Measure-
ment of Participate Adenosine-5’-Triphosphate. Handbook of methods in aquatic
microbial ecology, 359.
Knight-Jones, E. W. (1951). Preliminary studies of nanoplankton and ultraplankton
systematics and abundance by a quantitative culture method. Journal du Conseil,
17(2), 140-155.
Kolber, Z. S., Prášil, O., & Falkowski, P. G. (1998). Measurements of variable chlo-
rophyll fluorescence using fast repetition rate techniques: defining methodology and
experimental protocols. Biochimica et Biophysica Acta (BBA)-Bioenergetics,
1367(1), 88-106.
Kromkamp, J. C., & Forster, R. M. (2003). The use of variable fluorescence meas-
urements in aquatic ecosystems: differences between multiple and single turnover
measuring protocols and suggested terminology. European Journal of Phycology,
38(2), 103-112.
Lebaron, P., Catala, P., & Parthuisot, N. (1998). Effectiveness of SYTOX Green
stain for bacterial viability assessment. Applied and environmental microbiology,
64(7), 2697-2700.
Trafi Publications 10-2017
45
Li, F., Mahon, A. R., Barnes, M. A., Feder, J., Lodge, D. M., Hwang, C. T., Schafer,
R., Ruggiero, S. T., & Tanner, C. E. (2011). Quantitative and rapid DNA detection
by laser transmission spectroscopy. PLoS One, 6(12), e29224.
Li, F., Schafer, R., Hwang, C. T., Tanner, C. E., & Ruggiero, S. T. (2010). High-pre-
cision sizing of nanoparticles by laser transmission spectroscopy. Applied optics,
49(34), 6602-6611.
Li, W. K. W., Jellett, J. F., & Dickie, P. M. (1995). DNA distributions in planktonic
bacteria stained with TOTO or TO-PRO. Limnology and Oceanography, 40(8),
1485-1495.
Lomas, M. W., Bronk, D. A., & van den Engh, G. (2011). Use of flow cytometry to
measure biogeochemical rates and processes in the ocean. Annual review of marine
science, 3, 537-566.
MacIntyre, H. L., & Cullen, J. J. (2016). Classification of phytoplankton cells as live
or dead using the vital stains fluorescein diacetate and 5-chloromethylfluorescein di-
acetate. Journal of phycology.
Maranda, L., Cox, A. M., Campbell, R. G., & Smith, D. C. (2013). Chlorine dioxide
as a treatment for ballast water to control invasive species: shipboard testing. Marine
pollution bulletin, 75(1), 76-89.
Martínez, L. F., Mahamud, M. M., Lavín, A. G., & Bueno, J. L. (2012). Evolution of
phytoplankton cultures after ultraviolet light treatment. Marine pollution bulletin,
64(3), 556-562.
Maurer, B. N. (2013). Optimization and characterization of the bulk FDA viability
assay to quantify living planktonic biomass (Doctoral dissertation, California State
University Monterey Bay).
MEPC (2008). GUIDELINES FOR BALLAST WATER SAMPLING (G2). Annex
3, RESOLUTION MEPC.173(58). THE MARINE ENVIRONMENT PROTEC-
TION COMMITTEE.
Miller, A. W., Frazier, M., Smith, G. E., Perry, E. S., Ruiz, G. M., & Tamburri, M.
N. (2011). Enumerating sparse organisms in ships’ ballast water: why counting to 10
is not so easy. Environmental science & technology, 45(8), 3539-3546.
Nakata, A., Fushida, S., Matsuda, M., & Fukuyo, Y. (2014). Development of New
Method for Estimating Number of Viable Organisms in Ballast Water. Journal of
the JIME, 49(4), 512-517.
New Horizons Diagnostics. (2011). SMART ™ II Cholera O1. New Horizon Diag-
nostics Corporation, USA. Available from: http://www.nhdiag.com/pdf/Chol-
era_O1_Smart_II_insert.pdf
Olenina, I., Hajdu, S., Edler, L., Andersson, A., Wasmund, N., Busch, S., Göbel, J.,
Gromisz, S., Huseby, S., Huttunen, M., Jaanus, A., Kokkonen, P., Ledaine, I. and
Niemkiewicz, E. (2006). Biovolumes and size-classes of phytoplankton in the Baltic
Sea. HELCOM Balt Sea Environ. Proc. No. 106, 144pp.
Olsen, R. O., Hess-Erga, O. K., Larsen, A., Thuestad, G., Tobiesen, A., & Hoell, I.
A. (2015). Flow cytometric applicability to evaluate UV inactivation of phytoplank-
ton in marine water samples. Marine pollution bulletin, 96(1), 279-285.
Peeters, E., Nelis, H. J., & Coenye, T. (2008). Comparison of multiple methods for
quantification of microbial biofilms grown in microtiter plates. Journal of microbio-
logical methods, 72(2), 157-165.
Trafi Publications 10-2017
46
Penru, Y., Guastalli, A. R., Esplugas, S., & Baig, S. (2012). Application of UV and
UV/H 2 O 2 to seawater: disinfection and natural organic matter removal. Journal of
Photochemistry and Photobiology A: Chemistry, 233, 40-45.
Peperzak, L., & Brussaard, C. P. (2011). FLOW CYTOMETRIC APPLICABILITY
OF FLUORESCENT VITALITY PROBES ON PHYTOPLANKTON1. Journal of
phycology, 47(3), 692-702.
Pereira, N. N., Colombo, F. B., Chávez, M. I. A., Brinati, H. L., & Carreño, M. N. P.
(2016). Challenges to implementing a ballast water remote monitoring system.
Ocean & Coastal Management, 131, 25-38.
Picot, J., Guerin, C. L., Le Van Kim, C., & Boulanger, C. M. (2012). Flow cytome-
try: retrospective, fundamentals and recent instrumentation. Cytotechnology, 64(2),
109-130.
Reavie, E. D., Cangelosi, A. A., & Allinger, L. E. (2010). Assessing ballast water
treatments: Evaluation of viability methods for ambient freshwater microplankton
assemblages. Journal of Great Lakes Research, 36(3), 540-547.
Rotman, B., & Papermaster, B. W. (1966). Membrane properties of living mamma-
lian cells as studied by enzymatic hydrolysis of fluorogenic esters. Proceedings of
the National Academy of Sciences, 55(1), 134-141.
Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn, G. T., Erlich, H. A., &
Arnheim, N. (1985). Enzymatic amplification of b-globin genomic sequences and
restriction site analysis for diagnosis of sickle cell anemia. Science, 230(4732),
1350-1354.
Schreiber, U., Quayle, P., Schmidt, S., Escher, B. I., & Mueller, J. F. (2007). Meth-
odology and evaluation of a highly sensitive algae toxicity test based on multiwell
chlorophyll fluorescence imaging. Biosensors and Bioelectronics, 22(11), 2554-
2563.
Schreiber, U. (2004). Pulse-amplitude-modulation (PAM) fluorometry and satura-
tion pulse method: an overview. In Chlorophyll a Fluorescence (pp. 279-319).
Springer Netherlands.
Senapati, S., Mahon, A. R., Gordon, J., Nowak, C., Sengupta, S., Powell, T. H.,
Feder, J., Lodge, D. M. & Chang, H. C. (2009). Rapid on-chip genetic detection mi-
crofluidic platform for real world applications. Biomicrofluidics, 3(2), 022407.
SGS (2015). SGS ballast water training. Available from: www.sgs.com.
Shannon, T., Hatch, W. I., & Fitt, W. K. (2009). Evidence of photosynthate translo-
cation in an algal-acoel symbiotic system: An in vivo, qualitative approach. Journal
of Experimental Marine Biology and Ecology, 382(1), 69-75.
Shapiro, H. M. (1983). Multistation multiparameter flow cytometry: a critical review
and rationale. Cytometry, 3(4), 227-243.
Song, Y., Li, M., Yang, J., Wang, J., Pan, X., Sun, Y., & Li, D. (2014). Capacitive
detection of living microalgae in a microfluidic chip. Sensors and Actuators B:
Chemical, 194, 164-172.
Song, Y., Wang, J., Yang, J., Wu, Y., Li, N., Gong, N., Pan, X., Sun, Y. & Li, D.
(2012). Algae detection and ship's ballast water analysis by a microfluidic lab-on-
chip device. Instrumentation Science & Technology, 40(4), 305-315.
Stehouwer, P. P., Buma, A., & Peperzak, L. (2015). A comparison of six different
ballast water treatment systems based on UV radiation, electrochlorination and chlo-
rine dioxide. Environmental technology, 36(16), 2094-2104.
Trafi Publications 10-2017
47
Stehouwer, P. P., Fuhr, F., & Veldhuis, M. (2010). 2. A novel approach to determine
ballast water vitality and viability after treatment. Effects of various ballast water
treatment methods on the survival of phytoplankton and bacteria, 13. Emerging Bal-
last Water Management Systems, Proceedings of the IMO-WMU Research and De-
velopment Forum, Malmö, Sweden, pp. 233–240.
Stehouwer, P. P., Liebich, V., & Peperzak, L. (2013). Flow cytometry, microscopy,
and DNA analysis as complementary phytoplankton screening methods in ballast
water treatment studies. Journal of applied phycology, 25(4), 1047-1053.
Steinberg, M. K., First, M. R., Lemieux, E. J., Drake, L. A., Nelson, B. N., Kulis, D.
M., Anderson, D. M., Welschmeyer, N. A. & Herring, P. R. (2012). Comparison of
techniques used to count single-celled viable phytoplankton. Journal of Applied
Phycology, 24(4), 751-758.
Steinberg, M. K., Lemieux, E. J., & Drake, L. A. (2011). Determining the viability
of marine protists using a combination of vital, fluorescent stains. Marine biology,
158(6), 1431-1437.
Sudhaharan, T., & Reddy, A. R. (2000). Metal ion mediated inhibition of firefly bio-
luminescence: a possibility via a quaternary complex. INDIAN JOURNAL OF BIO-
CHEMISTRY AND BIOPHYSICS, 37(4), 256-267.
Suggett, D. J., Moore, C. M., Oxborough, K., & Geider, R. J. (2006). Fast Repetition
Rate (FRR) Chlorophyll a Fluorescence Induction Measurements.
Suggett, D., Kraay, G., Holligan, P., Davey, M., Aiken, J., & Geider, R. (2001). As-
sessment of photosynthesis in a spring cyanobacterial bloom by use of a fast repeti-
tion rate fluorometer. Limnology and Oceanography, 46(4), 802-810.
Throndsen, J. (1978). The dilution-culture method. Phytoplankton manual, 218-224.
Tobiesen, A. D., Bomo, A. M., Delacroix, S., Wennberg, A. C., & Liltved, H.
(2011). Limitations with respect to vital staining techniques for use in treated ballast
water. BALLAST WATER MANAGEMENT SYSTEMS, 13.
van der Star, I., Liebich, V., & Stehouwer, P. P. (2011). The forgotten fraction: The
importance of organisms smaller than 10 µm when evaluating ballast water treat-
ment systems. Ballast Water Management Systems, 41.
van Slooten, C., Wijers, T., Buma, A. G., & Peperzak, L. (2015). Development and
testing of a rapid, sensitive ATP assay to detect living organisms in ballast water.
Journal of Applied Phycology, 27(6), 2299-2312.
Veldhuis, M. J., Fuhr, F., Boon, J. P., & Ten Hallers-Tjabbers, C. C. (2006). Treat-
ment of ballast water; how to test a system with a modular concept?. Environmental
technology, 27(8), 909-921.
Veldhuis, M. J., & Kraay, G. W. (2000). Application of flow cytometry in marine
phytoplankton research: current applications and future perspectives. Scientia Ma-
rina, 64(2), 121-134.
Veldhuis, M. J., Kraay, G. W., & Timmermans, K. R. (2001). Cell death in phyto-
plankton: correlation between changes in membrane permeability, photosynthetic
activity, pigmentation and growth. European Journal of Phycology, 36(02), 167-
177.
Vilkner, T., Janasek, D., & Manz, A. (2004). Micro total analysis systems. Recent
developments. Analytical chemistry, 76(12), 3373-3386.
Trafi Publications 10-2017
48
Villareal, T. A. (2004). SINGLE-CELL PULSE AMPLITUDE MODULATION
FLUORESCENCE MEASUREMENTS OF THE GIANT DIATOM ETH-
MODISCUS (BACILLARIOPHYCEAE) 1. Journal of phycology, 40(6), 1052-
1061.
Villaverde, A., Guerrero, R., & Barbe, J. (1986). ATP Production after ultraviolet
irradiation inEscherichia coli. Current Microbiology, 14(1), 31-34.
Vitek. (2016) VITEK® 2: Healthcare. BioMerieux, Inc. Available from:
http://www.biomerieux-usa.com/clinical/vitek-2-healthcare
Wadsö, I. (2002). Isothermal microcalorimetry in applied biology. Thermochimica
Acta, 394(1), 305-311.
Waite, T. D., Kazumi, J., Lane, P. V. Z., Farmer, L. L., Smith, S. G., Smith, S. L.,
Hitchcock, G. & Capo, T. R. (2003). Removal of natural populations of marine
plankton by a large-scale ballast water treatment system. Marine ecology progress
series, 258, 51-63.
Wang, J., Sun, J., Song, Y., Xu, Y., Pan, X., Sun, Y., & Li, D. (2013). A label-free
microfluidic biosensor for activity detection of single microalgae cells based on
chlorophyll fluorescence. Sensors, 13(12), 16075-16089.
Weisberg, S. B., Noble, R. T. & Griffith, J. F. (2007). Microbial indicators of marine
recreational water quality. Manual of environmental microbiology, (Ed. 3), 280-289.
Welschmeyer, N. & Kuo, J. (2016). Analysis of Adenosine Triphosphate (ATP) as a
rapid, quantitative compliance test for ships’ ballast water. Moss Landing Marine
Laboratories. Report number: :HSCG32-13-R-R00016. United States Coast Guard,
New London.
Welschmeyer, N. A., & Maurer, B. (2011, October). A portable, sensitive plankton
viability assay for IMO shipboard ballast water compliance testing. In Proceedings
of the IMO globallast conference: compliance monitoring and enforcement, Istan-
bul, Turkey (pp. 127-139).
White, A. J., & Critchley, C. (1999). Rapid light curves: a new fluorescence method
to assess the state of the photosynthetic apparatus. Photosynthesis research, 59(1),
63-72.
Wier, T. P., Moser, C. S., Grant, J. F., First, M. R., Riley, S. C., Robbins-Wamsley,
S. H., & Drake, L. A. (2015). Sample port design for ballast water sampling: Refine-
ment of guidance regarding the isokinetic diameter. Marine pollution bulletin, 98(1),
148-155.
Wright, D. A. (2012). Logistics of compliance assessment and enforcement of the
2004 ballast water convention. Journal of Marine Engineering & Technology, 11(1),
17-23.
Wright, D. A., & Welschmeyer, N. A. (2015). Establishing benchmarks in compli-
ance assessment for the ballast water management convention by port state control.
Journal of Marine Engineering & Technology, 14(1), 9-18.
Wright, D. A., Welschmeyer, N. A., & Peperzak, L. (2015). Alternative, indirect
measures of ballast water treatment efficacy during a shipboard trial: a case study.
Journal of Marine Engineering & Technology, 14(1), 1-8.
Zetsche, E. M., & Meysman, F. J. (2012). Dead or alive? Viability assessment of mi-
cro-and mesoplankton. Journal of Plankton Research, 34(6), 493-509.
Trafi Publications 10-2017
49
Appendices
Appendix A: SGS ATP sampling approach
Trafi Publications 10-2017
50
Trafi Publications 10-2017
51
(SGS, 2015)
Trafi Publications 10-2017
52
Appendix B: FDA Pulse counting device
(Nakata et al., 2014)
Trafi Publications 10-2017
53
Appendix C: Indicative methods described in the present study
Table 8. Indicative methods explained in the present study (Throndsen, 1978, Bohren and Huffman, 2000,
Kromkamp and Forster, 2003, Li et al., 2011, Cangelosi, 2011, Fykse et al., 2012, Gollasch et al., 2012, Wang et
al., 2013, Song et al., 2014, Bradie, 2016).
Method
Description
ATP Measures luminescence in the presence of luciferase en-
zyme from seawater extraction.
FDA Measures fluorescence from living or dead cells.
PAM Measures photosynthetic activity and phytoplankton bio-
mass. Analyses living cells based on variable fluorescence
of chlorophyll of living algae.
Microscopy