This work is based on the conclusions and recommendations following the pre-study on ballast water carried out by DNV in 1999; Ballast Water Transfer Atlas (BWTA), Hazard Assessment and Decision Support Pre-study. This was a feasibility study evaluating the risk-based ballast water management approach, based on biogeographical qualities and generally accepted methods of assessing risks.

The EMBLA Integration phase has outlined mismatches between current understanding of the driving mechanisms associated with transfer of unwanted aquatic organisms and data availability. Associated topics related to regulation-development and the integration of a global management concept have also been addressed.

The development of EMBLA is proposed to be undertaken in three stages. The first (EMBLA 2000) aims to include an upgraded proposal to EU fifth framework programme. Regardless of the outcome of this, EMBLA 2000 will focus on the following main items:

• Data-processing and analysis (DPA)
• Standards and norms (SAN)
• Consequence assessment methodologies (CAM)

The required Infrastructure and user interface (IUI) of EMBLA requires input from the above items and is not given priority in the first development stage. This will not be a subject in the EU proposal either.

In addition to the items outlined, EMBLA Integration phase has recognised the need to gain acceptance of the principles upon which it is based among national maritime authorities. The project will therefore, in addition to the identified tasks of EMBLA 2000 outlined above, include an activity aimed at presenting and gaining international acceptance and recognition. "Target authorities" will be those who already take precautions against biological invasions, and those who presently support the EMBLA initiative.

To promote the EMBLA initiative, a web site, as well as a Ballast Water Library have been established.

The earth's biological diversity is being destroyed at a rate unprecedented in human history. The legally binding Convention on Biological Diversity (CBD) was the first international agreement obligating governments to conserve and sustain their biological resources and ensure the equitable sharing of their benefits. The Convention was set up for signature on 5 June 1992 at the United Nations Conference on Environment and Development (the Rio "Earth Summit"). It remained open for signature until 4 June 1993, by which time it had received 168 signatures. The Convention entered into force on 29 December 1993.

Most of the world's governments gathered in Jakarta, Indonesia in November 1995 for the second meeting of the Conference of the Parties (COP-2) to the CBD. The meeting marked for the first time, the international community’s comprehensive address to urgent global problems of marine and coastal biodiversity loss. The decisions taken on this topic were referred to collectively in the Ministerial Statement issued at COP-2 as the Jakarta Mandate on Marine and Coastal Biodiversity (Jakarta Mandate). These decisions can be found in two documents: UNEP/CBD/COP/2/19 and UNEP/CBD/COP/2/5.

At the national level, Parties are required to develop and implement comprehensive national biodiversity strategies and action plans. This includes, for example, to identify, research and monitor their biodiversity; establish protected areas; regulate or manage activities with significant adverse effects on biodiversity; and conduct assessments of the biodiversity impacts of proposed projects.

At an international level, institutional mechanisms are set up under the CBD to assist the Parties in their implementation efforts. This includes a financial mechanism to provide assistance to developing countries; an information clearing-house and an advisory body on scientific and technical matters (the Subsidiary Body on Scientific, Technical and Technological Advice, or SBSTTA). In addition, the COP which comprises all governments which have ratified the Convention, serves as the CBD's governing body, and currently meets once a year to review and strengthen CBD implementation.

In supporting the SBSTTA recommendations, the COP in effect recommended a "checklist" of actions that Parties should take to fulfil their obligations under the CBD in marine and coastal environments. These recommendations cover, in particular, five thematic areas:

Several of these thematic areas include to a lesser or greater extent elements of the "ballast water problem". However, the thematic area of "alien species" is particularly interesting in this context.

Integrated Marine and Coastal Area Management (IAM) involves an integrated, cross-sectored, and co-ordinated approach to managing marine and coastal resources, involving a wide variety of stakeholders and addressing a wide range of activities and threats.

Marine and Coastal Protected Areas (MPAs). MPAs are areas that have been designated for protection, in which activities and uses of the resources in the area are regulated to achieve conservation objectives. Many different categories of MPAs exist, based on the type and level of restrictions on resource uses and activities.

Sustainable Use of Coastal and Marine Living Resources (CMLR). The Over-exploitation of CMLR primarily resulting from industrial-scale, commercial fisheries and non-selective and destructive fishing gear and practices, poses a major threat to marine and coastal biodiversity.

Mariculture. Mariculture is the controlled cultivation in the sea of marine organisms in tanks, ponds, cages, and other structures. Mariculture on an industrial scale may pose several threats to marine and coastal biodiversity.

Alien Species. Alien species are species transferred by human activity, either intentionally or unintentionally, to an area in which they do not naturally occur. As noted by the SBSTTA: "Alien species have the potential for significant, non-reversible, adverse impacts on marine and coastal biodiversity. Such impacts generally tend to be unpredictable and tend to homogenise and simplify biotic communities. Eradication of established alien species is difficult, if not impossible".

Changing ballast water in open waters either by re-ballasting or by the flush-through methods as proposed in the existing IMO Guidelines (Ass. Res. A868.20), and as requested by a number of flag states prior to entering their national waters, does not provide adequate protection against the introduction of unwanted harmful aquatic organisms. The method of changing ballast water including the third option, that of dilution, suffers from restraints associated with, among other things, remaining vessel safety. Furthermore, even where application is possible, there is strong evidence which suggests that the efficiency of the method of changing ballast water is highly variable, and dependent upon factors related to each case in particular. In most cases, these circumstances are such that they cannot be manipulated, modified or altered. Hence, the current policy on ballast water measures should be considered only as an intermediate action to be adopted when applicable from a safety point of view. The protection it might provide is uncertain and most likely not sufficient in cases where a real risk of unwanted biological invasion is present.

By looking at the geographical patterns of typical "ballast-dependent" trades, it becomes evident that not all of these will represent the same element of risk associated with causing an unwanted transfer. Furthermore, the nature of the risk element will differ and hence the characteristics in relation to the most efficient preventive measure might equally differ. By assessing the actual risks (in terms of bioinvasion) associated with an actual ballast voyage based on the requirements of the presence of basic parameters, safe ballast voyages can be identified. Moreover, by considering the nature of voyages that might represent such a risk, requirements to preventive or risk-reducing measures might be identified.

Alternative preventive actions or measures are required to sufficiently secure against unwanted biological invaders. Such alternatives will have different effects on different species and are therefore likely to have specific areas of application. Hence, standards or norms for such measures are required. The development of a managerial system for assessing case-specific risks and identifying adequate actions or measures to be undertaken, should comply with such norms/ standards. Further, such a system should also include the ability to assess the potential of both an invasion as well as that of introducing a measure. A consequential analysis such as this should include aspects related to both ship safety and ecological damage as well as to economical losses.

Realisation of the BWTA concept as described in the pre-study undertaken by DNV in 1999, was proposed for the EU’s fifth Framework Programme. The proposal was submitted to the Programme on Energy, Environment and Sustainable Development – Sustainable Marine Ecosystems, first call, June 15. 1999. The project proposal was titled "Environmental Risk Management System for Ballast Water Assessment" and given the acronym EMBLA.

The proposal consists of three parts as required:

• Part A:General information on applicant and partners.
• Part B:Detailed description of work to be undertaken presented in work packages with references to performing institutions, time schedules and related costs.
• Part C:Description of contribution to the EU community represented by the proposed project.

Part B of the proposal is attached to this report (chapter 11).

The EU initiative was proposed by DNV. Taken into account the diversity of the project, a partner consortium was established to represent all the stakeholders involved. The proposal was forwarded on behalf of the following partner institutions.:

• Registro Italiano Navale (RINA), Italy. Contact Mario Dogliani
• Rotterdam Municipal Port Management, Netherlands. Contact Cornelis de Keijzer
• SSPA Maritime Consulting, Sweden. Contact Claes Källström
• University für Meehreskunde, Kiel Germany. Contact Stephan Gollash
• Åbo Akademi University, Finland. Contact Erkki Leppakoski

Following the EU screening and evaluation process, the following statement was received.:

"The proposal is a useful effort to compile data on the transfer of unwanted organisms by ballast water, which is a significant problem in the management of the marine environment. However, such transfer is also taking place in other ways, e.g. by fouling of ship hulls or together with shipping of living organisms such as oysters and mussels.

We suggest that the proposal should be submitted to a different programme."

The BWTA project group is of the opinion that the true objective of the proposal was not sufficiently addressed and hence not fully appreciated by the evaluator. The reasons for this are thought not to be arbitrary, but to also be related to the priorities of the proposal and will be a subject of the process of preparing a follow-up. This follow-up proposal will be based on the objectives of the original proposal but will in addition reflect on findings, conclusions and priorities made by the work presented in this report.

A revised project proposal is expected submitted in the beginning of year 2000.

The acronym of the EU proposal, EMBLA, has later been adopted as the formal notation for the BWTA concept as it has been developed. It should therefore be noted that following references to EMBLA refer to the BWTA concept as described in the DNV pre-study of 1999.

The development work of EMBLA is undertaken by a project team based at DNV’s headquarters at Høvik, Norway. The team has wide access to expertise in specific areas throughout the DNV network as well as through co-operating partners.

The project has communicated widely with different organisations connected to the issues involved. The reference group established during the Pre-study (DNV 1999) has also been active through the integration phase. Partners have contributed on an individual basis. Project progress has been communicated to the group through memos from the internal project progress meetings (5).

In preparing the EU proposal, a consortium was formed. These partners have been providing important input not only used in the proposal but also utilised in the work reported on here.

Literature and references to ballast water management and related issues are comprehensive, but not always easily surveyable. To ensure an adequate and easily accessible flow of information, a literature Library including references used and collected for the project development was established. At present, the Library exists as an Excel worksheet. The implementation of this into a more suitable database structure is intended.

As part of the Integration phase, an EMBLA web site has also been established. The site encompasses an open section as well as one with access restrictions. The latter is only available for the project team. The site can be found at:

EMBLA has been presented to both national and international groups, institutions, etc. Some of these presentations can be found at the web site. Further, this work has extended the development of the demonstrator that was established in the pre-study. This is also available from the site.

EMBLA is a ballast water decision-support management system aiming at gaining worldwide recognition and acceptance. The system will provide case-specific assessments of ballast voyages (DNV 1999). EMBLA will identify those voyages which do not represent an unacceptable risk of the transfer of harmful aquatic organisms. These voyages can then proceed without delay. For some voyages, the system will identify parameters providing an element of unacceptable risk. The cause of these risks will be "isolated" and the system will provide a statement clarifying the circumstances. The system will further provide advisory as to actions to be taken.

A number of flag-states (14) have already implemented precautionary measures to vessels entering their national waters in ballast condition. Some nations strongly advise vessels to carry out ballast water exchange prior to entering, whilst some merely require a report of ballast water, amount and origin. At present there are only some few "local" mandatory requirements associated to ballast water precautions.

A future ballast water regulative regime will require a management system ensuring that potential biological threats are identified and that appropriate action is initiated. EMBLA will respond to such requirements. The users of EMBLA can be both port-states, as well as ship owners.

Shipping can be sectored into two main categories, firstly scheduled routes and that of non-scheduled. The latter will require that ballast water support by EMBLA is available on a 24/7 basis and can perform rapid assessments. EMBLA is therefore developed highly automised and accessible via Internet.

For fixed scheduled vessels, pre-assessments can be provided which have a limited validity. EMBLA will store all past assessments, and in doing so build up a ballast inventory for the specific vessel.

Table 3.4 below contains definitions of wordings used in the summarised report.

An ecological revolution in the biodiversity of the shallow waters of the ocean - the waters that we rely on for fisheries, industries and for recreation is now happening (Carlton 1993). Alien animals and plants are invading estuaries, bays, rocky shores and other coastal ecosystems at a rate of decades and years. What nature took millions of year to create, human activities are homogenising in a few decades through transport mechanisms which constantly cross the natural barriers of open oceans and continents. The introduction of alien species by human intervention is not new; ships have moved marine animals and plants for centuries. However, controlling coastal environment destruction and alteration, and eliminating alien species dispersal vectors will predictably lead to fewer and fewer invasions in the twenty first century.

Due to increased attention, the field of coastal rehabilitation has been steadily growing over the last two decades. According to Spurgeon (1999), coastal habitats provide a vast array of benefits to mankind in the form of goods (products) and services (functions). Since few of the goods and services are traded in the market place, they rarely have a readily observable monetary or financial value. However, they can have a considerable socio-economic value, particularly when utilised on a sustainable basis.

Living organisms in coastal water are loaded into ballast tanks along with the water when a vessel is ballasting. If a ship takes on ballast water in a shallow area, sediments and any associated organisms may also enter the ballast tanks. The release of ballast water may introduce non-native or alien species into the port of discharge. Once in a new environment, an organism may simply die, or it may take hold and reproduce, but with little noticeable effect on its surroundings. However, introduced species sometimes spread unimpeded with devastating ecological and/or economic results. Some well-known examples are listed in DNV’s Pre-study (DNV 1999).

Introduced species are those that have been transported beyond their natural range. Instead of being members of an ecosystem developed over time, these animals were transported beyond barriers that defined their natural range. The new habitat challenges the organism to perish or persist. An introduced species can spread rapidly, for instance where their natural predators, competitors and pathogens are absent from their new environment.

The role and impact of alien species in the marine environment is not well documented. Typically, very few organisms are able to survive in new surroundings because temperature, food and salinity are less than optimal. However, the few that survive and establish a population have the potential to cause changes in the receiving ecosystem. The impacts can be divided into two areas; ecological and economic, these two areas are however interdependent. Most of the observed effects have been detrimental and irreparable, displacing native species, and altering trophic level structure. Occasionally, alien species reproduce with natives and produce hybrids. Hybrids change the gene pool in an area and can simplify the ecosystem.

There are two basic approaches in dealing with bioinvaders:

• stop them from invading in the first place, or
• eliminate organisms that have invaded.

Eradication of established invaders is practically impossible. Preventing invasions from occurring is the more practical and economical solution in the long term.

No ballast water treatment method can so far completely eliminate the risk of introducing alien species. The goal should therefore be to build up a ballast water management system that minimises the risk of species introduction.

In order to predict the risk for successful establishment of an alien species in a new ecosystem/environment, basic information about species known to be invaders together with their natural history, community structure, and the biodiversity of their regional system is essential. The use of biogeographical regions and hazard species lists is therefore a main input in a risk assessment (described in chapter 8), estimating the risk for successful transfer of hazard alien species and their establishment in a new environment.

The acknowledged work of Briggs (1974) and Ekman (1953) constitutes the basis for the proposed marine biogeographical regions used in EMBLA. Their works provide a general and global distribution of the regions and are therefore preferred as a basis for the system. In cases where individual nations can verify the need for more detailed region division, this should be integrated in the system.

The regions are based on zoogeographical and temperature barriers. A zoogeographic barrier will usually separate regions of different geographic areas and/or different climatic history. Consequently the degree of species diversity and hence ecosystem stability will also differ between regions. The region that develops the greatest ecosystem stability, will function as a distribution centre and supply species to less stable areas, but it will accept few or none species in return. Two areas with ecosystems of approximately equal stability can experience successful invasions in both directions. The risk for introduction of alien species increases in an ecologically unstable area. The biogeographical regions are described thoroughly in the pre-study (DNV 1999).

Barriers determine where organisms live and how natural ecosystems are created. Alien species live in places that they could not reach by natural dispersal (tides, currents etc.).

Physical barriers can be land, oceans (large distances, large depth), salinity and temperature which to a greater degree are absolute barriers. Other barriers such as unsuitable food and habitats, the presence of pathogens, predators and competitors, are not absolute barriers. Natural ecosystems are dynamic and the range of species changes over time, spreading or retracting as environmental conditions fluctuate and food availability, predators and diseases influence populations.

On commission for among others the World Bank, a global representative system of Marine Protected Areas (MPAs) (Kelleher et. al. 1995) has been established. The role of MPAs is briefly described in chapter 2. MPAs are a practical way of conserving marine biodiversity, maintaining the productivity of marine ecosystems as well as contributing to the economic and social welfare of human communities.

The criteria used to identify priority areas in the report of Kelleher and coworkers (1995) were developed by Kelleher and Kenchington (1992) and have been adopted by the International Maritime Organisation (IMO) for use in the identification of particularly sensitive areas. So far 1306 MPAs have been identified around the world. The main focus is on areas with a subtidal element.

• The MPAs (marine protected areas) are not yet incorporated in the system of regions, but it is essential that the existing and forthcoming MPAs are incorporated in the system as a part of the existing regions.
• The preservation of coastal biodiversity is an international aim (chapter 2). However, it is up to the recipient nation, to decide on the degree of immigration and further dispersion out of port. Some countries will doubtless be very conservative, whilst others will focus on avoiding species known to lead to economical problems.
• The probability for survival of a species in a port area could differ from that of a species outside the port area. The environmental parameters in the port can support a local establishment for a species, while the surrounding waters do not support establishment and vice versa. It is therefore essential to include the environmental conditions (salinity, habitat etc.) of the specific port in the risk matrix.
• Scientific knowledge of the natural ecosystem of a port and its surrounding waters will differ greatly; different areas of the world have been investigated at differing levels and to varying degrees, mirroring the community structure and biodiversity.
• Modified or disturbed ecosystems, for instance a port, will be vulnerable to the establishment of new species, regardless of the general ecological stability in the surrounding area. Immigrating species (often from another port) with flexible living requirements and high reproductive rates could be highly invasive in the absence of natural competitors and predators in the recipient areas.
• In many cases, the shipboard ballast water will contain a mix of water and sediment from multiple ports. This factor could make it necessary to include back tracing ballast water history.

Literature concerning the topic of aquatic organism transfer has identified a need to "rate" certain species with a link to some geographical origin. The terms "target lists" and "hazard lists" are frequently used.

Some nations have or will create special country or port specific target lists; i.e. lists of unwanted organisms. EMBLA has established a target list concept, which could be used in this context. The target list could range from species known to cause economical or ecological harm to any species not naturally occurring in the coastal region of the country.

EMBLA will prepare a list based on known historical transfer of invasive species. The information in this list gives central information to the risk assessment (see chapter 7). The list also gives information on the region of origin and which regions the species are distributed in. Other information such as rate of spread, salinity- and temperature tolerances, type of habitat, known ecological-, economic- and aesthetic effects are also included in the list.

The list will require regular updating and must be combined with a list of environmental variables for each port. The information in the list combined with environmental data in the recipient port is essential for estimating the risk of establishment of an invading species.

• A problem, which often will arise, is that many countries possess insufficient knowledge on national marine ecology, to predict the likelihood of certain species being able to cause major effects. In these cases a target list could be constructed from the hazard species list.
• Available information of nonindigenous species is widely scattered, sometimes obscure and highly variable in quality.
• No institution/ mechanism exists for monitoring new invasive species on an international basis.
• The effects of many known invasive species have never been studied and information on effects is similarly lacking for undetected species, which are already is established.
• Relatively few invasive species cause great harm, but there are time lags and unknown effects. For example, changes in environmental conditions could lead to some species eventually causing great harm.
• Lack of environmental data in ports.
• Predicting the role and effects of a species in a new environment is extremely difficult. Each introduction creates a new combination of organism and environment. Detailed information on both is necessary to anticipate the result. Such information is usually lacking.
• A new environment may affect rates of reproduction, susceptibility to disease and other features that affect a species’ success in becoming established.
• The effects of some species could change as they enter new environments - a factor making predictions difficult.
• Species characterised as cosmopolitan are often highly tolerant to different environmental variables and are assumed to be living in all regions of the world. An invasion of such species into a port could therefore not necessarily be solely connected to discharge of ballast water. Such species are also likely to move in to the port from the surrounding waters.
• Identification problems of invasive species could make it necessary to include all species in a genus as possible invaders. However, correct identification is vital for distinguishing invading species from native species and for establishing a management programme.

Each day more than 3-4000 species are being transported around the world in the ballast water and sediments of ships (Carlton & Geller 1993, Carlton et al. 1995, Gollasch 1996). These organisms include everything from phytoplankton, zooplankton, virus and bacteria, to a wide variety of invertebrates, and small fish. Most of the higher taxa are represented, and the animal taxa most commonly found are copepods, barnacle larvae, nematodes, rotifers, polychaetes and cladocerans (Carlton 1985). This implies that a number of organisms are being shipped to areas where they are non-indigenous.

Sampling of the ballast water is essential in order to establish whether vessels are transporting non-indigenous species in their ballast tanks or not. If any kind of transfer prevention technique is being implemented on board, sampling can also verify the success of the treatment.

The literature on sampling techniques has been assessed (table 5.1). Currently used sampling methods has been identified and their quality/ reliability, feasibility, costs, and compatibility with other comparative methods has been investigated. The literature was obtained by searches, both in university libraries and at the library of DNV, as well as on the Internet. Documents and material received as handouts at conferences and meetings has also been used. It should be noted that during the search several obstacles were met. Some literature was difficult to obtain; some reports were only intended for internal use, and some seemed only to be distributed through conferences and workshops. Often, only abstracts of work were obtainable or results had not yet been published. On these grounds it is acknowledged that important works may be absent from this literature study. It must also be pointed out that this literature study covered all aspects of the ballast water problem, and all the literature references and homepages were listed in a file. This is available through the EMBLA Library.

Ballast tanks are for obvious reasons located in vessels’ bottom and sides (fig. 5.2). Onboard there may also be other water-filled containers/ devices that can contain organisms, for example temporarily empty cargo tanks, fire mains, bilge tanks, oily-water separators, holding tanks, and marine sanitation devices (see also below). In addition to ballast water and sediments, ships also carry fouling organisms on the hull, thereby introducing organisms by three different means during one trip. Typical fouling organisms are bivalves, cirripedia, and tubeworms. This study however focuses on ballast water.

The first studies on ballast water sampling seems to have been undertaken in Australia (Medcof 1975) where the ship investigated yielded plankton of polychaetes, copepods, amphipods, ostracods, and chaetognaths. Several studies have since followed (see table 5.1). Some studies have revolved around specific groups, like dinoflagellates and diatoms (Hallegraeff & Bolch 1991, 1992), virus (McCarthy & Khambaty 1994), and protists (Galil & Hülsmann 1997). Taxa in the sediments have also been investigated (Kelly 1993). An IMO report from 1997 (Gollasch 1997) lists the results of a number of ballast water sampling studies to indicate the variety of species found, and Cohen (1998) compares the number of distinct taxa and densities of organisms reported in different studies from all over the world.

Several other sampling studies are in progress. The Smithsonian Environmental Research Centre (SERC) is conducting a project in Prince William Sound, Alaska and will have a report finished by the end of 1999. In England and Wales a project is being conducted on behalf of The Ministry of Agriculture, Fisheries and Food (MAFF) and it is in co-operation with Scotland (ICES 1999). In the Netherlands, AquaSense (AquaSense 1998) is undertaking a study with emphasis on sampling of phytoplankton and zooplankton. An international network for marine invasion research, INFORMIR, is also established to ease the exchange of information and co-operation between researchers (Gollasch 1997).

The problem of sampling studies include the fact that no standardised method has been applied during sampling, leading to quite a number of different sampling designs (Gollasch et al. 1998). Water samples from ballast tanks can be obtained through horizontal or vertical manhole covers or sounding pipes, from pipes during de-ballasting, or from water remains after de-ballasting. Specimen can be collected from these by means of water samplers (Ruttner, Van Dorn), buckets, pumps, or traps, or with plankton nets with a variety of different mesh sizes, opening diameters and shapes, lengths, etc. In addition, samples can be collected at different levels of the water column, at different time-intervals, during shipping or at port, and before or after treatment of ballast water. Table 5.2 give a summary identifying sampling access while table 5.3 summarises combinations of sampling locations for ballast tanks on different vessel types.

A number of studies on ballast water sampling have been carried out and others are in progress world-wide. In spite of the lack of mutual norms or standards, the institutions engaged seem to put considerable efforts into distributing information on their work. However, details and results are not always easily accessible.

Germany is co-ordinating a major study; "Testing Monitoring Systems for Risk Assessment of Harmful Introductions by Ships to European Waters". This project has been in progress for approximately two years and was expected to report at end off 1999. This is an EU Concerted Action project funded by the EU with the IMO as a partner. Several European countries are active participants but other institutions are also involved in the work. One of the subject areas includes comparing and harmonising various sampling methods, and thereby studying their effectiveness. Several land- and ocean-based workshops are included for testing the methodologies, and ocean-based workshops have been completed, for example on vessels between St. Petersburg and Lisbon, and Cork (Ireland) and Sture (Norway) (Gollasch et al. 1998, Gollasch’ homepage, Gollasch 1999). Gollasch and his co-workers have also tested different methods using a land-based plankton tower in Helgoland with known densities of specimens (Gollasch et al. 1998). Results after the first of these intercalibration and ocean-going workshops were:

• different methods must be applied in accordance to tank accessibility
• manholes are preferred to sounding pipes
• net samples are preferred to pumps
• the large Monopump is more efficient than the small handpump, but cumbersome to use
• the short cone-shaped net is preferred to a larger net (for zooplankton samples)
• a Ruttner sampler is the best device for phytoplankton samples (Gollasch et al. 1998)

After the fourth land-based workshop in Wales 1999 it was recommended that the number of sampling methods be narrowed down to include only the most effective, so one tank could be sampled more than once a day (Gollasch 1999). For zooplankton the most efficient sampling devices were the 55 um cone net, 55 um net, bucket, and hand pump. (There were no conclusions for phytoplankton.)

The project is being completed in 1999, and the final report is soon to be published (Gollasch in prep.).

In Australia, a sampling program was found to be necessary as a component of the Australian Quarantine Inspection Services (AQIS) decision support system for the assessment of risk of introduced species (Sutton et al. 1998). This program was to include an evaluation of a selected group of methods regarding practicality and effectiveness, by contacting international groups working on sampling to identify methods, and testing the methods on vessels under a variety of conditions.

Sutton et al. (1998) collected information from researchers all over the world during their procedures regarding ballast water sampling. Of nine identified methods, seven were evaluated for their operational application and effectiveness in sampling zooplankton and a suite of target taxa. The seven methods were sounding pipes, manholes, air vents and breather pipes, in-line sampling, deck/fire taps, fixed position sampling, and sampling the port community. Fixed position sampling involves a fitting of hoses on the manhole hatches for sampling at specific "fixed" depths. Field trials of the methods were a large part of the study.

In the report the sampling locations and equipment (nets, pumps) are described, as are the details of the methods and respective research institutions.

Practicality and relative effectiveness were considered during the evaluation task. The effectiveness was found by comparing density and richness of taxa for each set of method comparisons. An overview of the evaluations by Sutton is listed in table 5.4. Table 5.5 shows an overview of the suitability for each method related to target taxa.

Conclusions of Sutton et al. (1998) is summarised below:

• Access to the tanks and the stage of ballast cycle limits the choice of methods available. It is advised not to rely on one single method because this will not account for all possible situations on the vessel or be able to sample all taxa.
• For phytoplankton and microbes, any tested method was adequate for monitoring or testing, but none were adequate for sediments.
• For zooplankton, there were significant differences between methods and all showed sampling bias. The bias can be caused by organisms’ attempting to avoid becoming caught by swimming, heterogeneity in tanks or comparisons of methods (between tanks, between vessels, seasons, age of ballast water, time between samples, exchanged ballast water, etc.). Sampling from manholes by nets was over all the most effective.

Both New Zealand and Australia have several activities related to the ballast issue. Ongoing work in New Zealand covering sampling is discussed in the following.

A manual for sampling sounding pipes

The Cawthron Institute in New Zealand has compiled a practical manual for sampling ballast tanks via ships’ sounding pipes (Dodgshun & Handley 1997, also in Hay et al. 1997). This manual examines two important sides of sampling, namely the preparation beforehand and the actual sampling process onboard. The work is extremely thorough. The method is chosen because of its simplicity, quickness, repeatability, minimum assistance requirement, disruption to routine, and because it can obtain samples from every tank on board which are fitted with a sounding pipe.

Instructions on other types of sampling are included for situations where it is impossible or unnecessary to sample via sounding pipes. For example when samples can be taken directly from the ballast tank using a plankton net where the impeller pump does not reach down deep enough in the tank or when the inertia pump hose only functions inside a narrow pipe. The sampling of ballast tanks with nets through the manholes also provides both quantitative and qualitative information (the quantity of water is known because of the diameter and length of nets, and can be related to the number of organisms retrieved).

The sampling comparisons

The practical manual for sampling of sounding pipes mentioned above (Dodgshun & Handley 1997) was part of a bigger project on ballast water (Hay et al. 1997). Another major part was to compare samples according to a number of variables such as type of tank and type of vessel, origin of water, possible ballast water exchanges (and their volume and frequency), and containment time.

In addition to investigating the phytoplankton and zooplankton in the tanks, the study compared the efficacy of the Waterra inertia pump & centrifugal petrol pump (impeller pump) and determined an optimum sample volume. The optimum sample volume was standardised to 100 litres based on comparisons of taxa in 50 litre volumes. For the pumps, the inertia pump sampled both more species and a greater abundance of species than the petrol pump. The centrifugal pump was seen to damage taxa.

When comparing taxa by ship type it was found that the number of living phytoplankton were greater in bulk carriers and break bulk carriers than in container ships. This also goes for the zooplankton, but these where found to be in about 10% more of the container ships.

For tank types and phytoplankton there was little evidence of differences in container ships, but the lowest numbers of phytoplankton were found in double bottom tanks and the highest numbers in lower wing and forepeak tanks. The number of invertebrates in container ships was also lowest in double bottom tanks, and the highest numbers were found in upper wing tanks.

With regard to bulk carriers, the majority of tanks contained phytoplankton, except some double bottom tanks, while the highest numbers of invertebrates was found in holds, forepeaks and upper wing tanks. The lowest numbers of phytoplankton were found in double bottoms and upper wings, and the lowest number of invertebrates in double bottom tanks.

One explanation for the distribution of plankton was that tanks lacking plankton were used less frequently. This goes especially/specifically for the double bottom tanks. It was underlined that the numbers of some of the sampled ship types were possibly too low for the comparisons to be accurate.

The methodology is compared to the method of sampling through manholes, and the disadvantages of each method are compared. As far as manhole sampling is concerned, the "lid" need to be removed and thus incurring assistance from the ships’ crew; cargo may be covering the manhole; it might be situated lower than the water-level or access can be through two or more other (possibly water-filled) tanks, and there could be a safety violation. For sounding pipes sampling there are two disadvantages. Some ships do not have sounding pipes, and sampling by this means is thereby impossible. The samples also only come from the bottom of the tank, which leads to the question as to whether the taxa in this level are representative for the water in the tank as a whole. To counteract this effect, it is suggested that plankton net samples could be taken from the top layers where possible.

For a discussion on the exchange of ballast water, see chapter 6.

The lack of standardisation for sampling, resulting in a great variety of sampling designs is clearly seen when comparing the methods. To reach sound conclusions from these comparisons is difficult. Each study has used their own method notwithstanding other works, and some have thereby sampled both qualitatively and quantitatively, while others have only sampled qualitatively. The last category can not standardise the number of taxa per unit of volume, but only give a number on organisms (e.g. Williams et al. 1988). The results of the studies are thereby not very comparable.

The works of Gollasch and co-workers, Sutton et al. (1998) and Hay et al. (1997) have tried to compare several different methods. The conclusion seems to be that sampling with nets through manholes is the most effective way to get good results. When there is no access to manholes, sounding-pipe sampling with pumps may be the best option.

The studies have utilised different types of pumps, and in comparison the pumps seem to be judged the following way: The monopump is more efficient than both the Waterra inertia pump and the small handpump, and the impeller pump is inferior to both the inertia pump and the monopump. The problem lies in the monopump’s limitations when the head (distance between water level and deck) is greater than 6 meters. Under these circumstances the monopump must be replaced by the inertia pump. The fact that this is a more cumbersome method than for example the impeller pump can not prevent it from being used. Instead, it must be noted that this is the most effective method.

It must be stressed that if sampling via manholes is by far the most efficient way of sampling, there is a need for relevant authorities to establish guidelines when sampling is to be done by scientists, quarantine agencies, or others. In that way the sampling personnel saves a lot of paperwork and the ship’s crew can take such precautions as may be necessary, like ensuring that manholes are accessible after loading. In this way, when entering a vessel, it will be possible for each of the sampling methods to be performed.

To get a clear picture of the diversity of organisms on board a ship, it is necessary to obtain samples that are representative of the ship as a whole. Representative samples mean that replicates from each tank must be collected, and from as many tanks as possible. Taking a couple of samples from one tank will not give a clear picture of the organisms being transported on a ship. Very often, studies give no information as to whether more than one tank has been sampled or not, which again give little information of particular value. Hay et al. (1997) are one of the few who compared tanks within ships, and they did find some evidence of differences between tanks, but tested too few ships to get an accurate picture.

Laboratory simulations where several sampling methods are applied on the same water masses are crucial for determining differences between methods. Compared to field trials, laboratory simulations could be the easiest way to get clear results, as the conditions are monitored in a controlled environment.

As also stressed by Sutton et al. (1998), the efficiency of the method applied largely depends on clearly specifying the aims of the study. Sampling can be applied in the routine monitoring of survival of organisms during every voyage (as would be the case for establishing the efficiency of treatment), alternately, sampling could be done on an individual basis (such as when scientists border a vessel to sample specific taxa). Some references assume that in the future sampling will be implemented on every vessel to verify treatment, and the method must thereby be simple, quick, and not disrupt the other activities onboard. With respect to this, the sounding pipe sampling described by Dodgshun & Handley (1997) would be most appropriate.

Regardless of methods chosen, some preparative actions onboard should be taken. This would aid sampling to be performed under controlled conditions, and the industry would have to recognise the need for sampling to be carried out.

Based on the literature used in this study and its incomparable findings, no clear conclusions towards a single method can be obtained. Additional studies in field and laboratories are needed before further assessments can be made.

The literature survey carried out has not provided a sufficient base for detailed methodology assessment, as the references of most studies did not give any reference to the methodology. This is also the case for the summaries of the studies.

If further literature on ballast water and introductions of species is needed, it would be advisable to start with the compiled literature-list in Cawthron’s Report No. 417 (Hay et al. 1997), and the multiple articles and reports of Dr. James T. Carlton. Gollasch (1997) lists multiple homepages and mailing-lists. These could be of help when investigating ballast water issues.

Considerable efforts have been put into developing new technologies or assessing the applicability of known methodologies in order to eliminate or eradicate organisms in ballast water.

Most technologies proposed or presented as a transfer prevention measure are often adopted from other applications than that of treatment of seawater.

In order to establish a base of understanding in relation to the issue of treatment options necessary when considering the issue of risk reducing measures, a literature assessment has been undertaken. The aim of this being to address the availability of such measures likely to enable reliable cost-efficient ballast water treatment.

Relevant literature was identified through the search described in chapter 5.1 and also include presentations, reports and documents received during meetings and at conferences. It should be noted that obstacles were experienced when approaching some sources, and some relevant studies may therefore not have been addressed.

Common to some of the studies identified, is the approach of summarising techniques in a prioritised manner that is attempting to visualise their individual feasibility as a function of some identified parameters. Typical parameters are those of efficiency, cost and safety. Other important characteristics are also focused upon.

The aim of the assessment undertaken here has been to compare existing proposed transfer prevention techniques based on literature findings with respect to:

• Feasibility
• Safety (occupational and environmental)
• Cost efficiency
• Quality/ certainty of results

Study reports and scientific articles identified, covering one or more prevention techniques are listed in table 6.1. Table 6.2 summarises the treatment options and describes their individual mode of operation. Table 6.6 provides further information on the prevention techniques assessed.

A recent report (Oemcke, D. 1999) has provided input to this chapter on a general basis. This work emphasises that until a system of comparing options on the basis of the cost of removing particular organisms to a pre-determined level of invasion risk is developed, it will be difficult to rationally compare ballast water treatment alternatives. This statement is supported by other works.

Some major studies have been identified:

• The American National Biological Invasion Shipping Study (NABISS) (Carlton et al. 1995) investigated 32 identified alternative technical prevention solutions and ranged them in accordance to efficiency, costs, feasibility, and safety, se table 6.3 (Carlton et al. 1995, also sited in AquaSense 1998).
• An IMO report (Gollasch 1997) describes 36 different options in general terms but emphasises cost aspects from a selection of these.
• The Committee of Ships’ Ballast Operations et al. (1996) have published a book, "Stemming the tide" which includes a large section on treatment prevention techniques.
• The Pollutech Environmental Ltd study (Laughton et al. 1992) reviews and evaluates treatment options for The Great Lakes.

Carlton, Reid & van Leeuwen (1995) included an analysis of control options for the coastal waters of the United States (except The Great Lakes, but this area is dealt with in Reid & Carlton 1997 as a part of a larger study). Their conclusion was that an integrated management system, and hence the choice of a number of alternatives, is the most effective approach. Thirty-two control options and alternatives were identified and evaluated. These are listed in table 6.3. Of these, 16 options where considered pursuable for further study. These are presented in table 6.4.

The management options were divided into 4 categories:

The criteria used to evaluate and analyse the identified options included those of human and vessel safety, costs, biological effectiveness in removing or killing organisms, shipboard operational reality, post-implementation monitoring and assessment, and environmental impacts.

Table 6.5 summarises some aspects concerning ballast water exchange addressed in the study.

In addition to informing about activities on ballast water around the world, and the results of several sampling studies, regulations and guidelines, Gollasch’ (1997) comprehensive IMO report indulges in describing different treatment options to remove or kill organisms in ballast water. The options are reviewed on the basis of effectiveness, practicality, cost, and environmental- health and safety aspects. Costs associated to the options are also included.

The work concludes by claiming that a combination of techniques are required in order to achieve significant results, since not one option alone fulfils all the necessary criteria. Generally, a mechanical removal is recommended, rather than a physical or chemical treatment. Treatment on board the ship is said to be preferred to land-based treatment, but the reasons for this are not explained. For on board-treatment, there are also two choices of venue:

• en route treatment
• treatment in port

An attempt to summarise the evaluations has been made in table 6.6.

The most promising combinations of treatment are identified as being filtration or heat followed by changing salinity. Physical techniques are said to be more acceptable than chemical options. Heat and UV radiation are chosen as the most promising due to limited side effects on the environment and to limited potential consequences associated to safety aspects. Ozonation is explicitly mentioned as a positive exception amongst the chemical options.

A brief summary of these are available in attachment (chapter 12).

The Committee of Ships’ Ballast Operations (1996) in the US confined the discussion to contain on-board treatment only. They focus on treatments that are safe, practicable, feasible, cost-effective, and environmentally acceptable. The most important evaluation criteria are safety and effectiveness. The most flexible option is said to be the ultimate solution. The Committee looks at one option at a time, and does not discuss the use of combinations of options.

The considered techniques were filtration systems, oxidising and non-oxidising biocides, thermal techniques, electric pulse and pulse plasma techniques, ultraviolet treatment, acoustic systems, magnetic fields, deoxygenation, biological techniques, and anti-fouling coatings. The options considered most promising are listed in table 6.7.

For further reading, se attachment, chapter 12.

The Pollutech study (Laughton et al. 1992) reviews alternative treatment, management and control measures for The Great Lakes, and evaluates their efficiency. The study is divided into two parts in which the first reviews all available options and characteristics that might influence their efficiency and the second evaluates all options.

The ballast water characterisation included both biotic and abiotic parameters. Biological parameters were for example types of organisms and tank habitats. Of the abiotic parameters there are both design characteristics and physical parameters, like function of ballast water, volume, and pumping rates (design), and temperature, light, and salinity (physical) etc.

The 14 alternatives listed are:

Options ruled out for further considerations were: increased length of voyage, electrical discharge, flue gas, and high-velocity sharing. These lacked information, or had questionable efficiency and/or problems with safety.

Seven evaluation criteria were weighed (see below) and each treatment option was rated (below average, average, above average) according to each evaluation criterion.

Occupational health and ship safety: 24 %
Environmental acceptability: 18 %
Control effectiveness: 16 %
External monitoring requirements: 12 %
Maintenance and operation: 12 %
Cost effectiveness: 10 %
Technical practicability and feasibility: 8 %

This resulted in a list of the options with highest through to lowest score giving the following results.:

• Ultrasonics and heat treatment were eliminated on the grounds of being neither practical nor effective.
• The silicone based anti-fouling paints scored low on both Costs and Maintenance and Operation, and were deemed impractical.
• The chemical treatments ranked lowest because of a below-average score for Environmental Acceptance and Occupational Health and Ship Safety, in addition to Maintenance and Operation.
• Physical methods ranked above average for filtration, UV light, shore facilities, certified ballast water, and non-release.
• The second highest ranking was given to discharge to sanitary sewer after on board treatment (e.g. filtration/ filtration + UV) or sanitary sewer handling followed by wastewater treatment plant.
• The most favourable option was a self-cleaning filtration unit screening which could filter organisms over 50 µm.
• The combination of UV and filtration could be more effective and thereby be rated highest, but the costs of the unit, maintenance and installing of the unit reduced this rating.

Concluding summary tables for some of the options are listed in chapter 12.

This study (AquaSense 1998) deals primarily with risk assessment of ballast water but also includes prevention techniques (risk reducing measures). Methods assessed included filtration, heat, ballast water exchange, and the Brazilian dilution technique. The information provided is largely collected from other references (addressed above). Table 6.8 illustrate some of the findings.

Gollasch (1997) consider the chemical options in general as less acceptable due to the by-products produced that are often environmentally unacceptable. Hypochlorite is mentioned as a mechanism that could be used in case of emergency, and metal ions such as copper and silver are worth further investigation. Adjusting salinity is said to be clearly useful when the right supplies are available.

The Committee on Ships’ Ballast Operations et al. (1996) considered the options ozonation, electric pulse and pulse-plasma techniques as limited applications. Options considered to be negatively were UV-light, acoustic systems, magnetic fields, deoxygenation, biological techniques and anti-fouling coatings. The first two options have limited effectiveness on higher organisms, while the effect of magnetic fields is unknown. Deoxygenation is said to be ineffective against anaerobic bacteria, and cyst and spore stages.

The Pollutech study (Laughton et al. 1992) did not give the options of increased length of voyage, electrical discharge, flue gas, and high velocity sharing further consideration. These lacked sufficient information, or had a questionable efficacy and/or problems with safety. Ultrasonics and heat treatment were eliminated on the grounds of not being practical or effective, and the silicone based anti-fouling paints scored low on the cost-effectiveness and maintenance and operation evaluation criteria, and were deemed impractical. The chemical treatments rated lowest because of a below-average score for environmental acceptance and occupational health and ship safety, in addition to maintenance and operation.

Exchanging ballast water during transport is most likely the only used method at present and represents a practical option in reducing the number of organisms in ballast water. The method is based on the assumptions that organisms taken on board a ship in coastal areas will not survive when released in oceanic areas, and vice versa, and that mid-ocean water contains relatively few species. IMO Guidelines recommends mid-ocean exchange, with mid-ocean being at least 200 nautical miles from the coast, at a minimum depth of 500 m (2000 m (Cohen 1998)), and with a salinity at or above 30 ‰ (Hay & Tanis 1998).

Ballast water is used for ensuring stability, adjustment of trim and heel, and as a replacement for cargo to keep the appropriate depth/ weight distribution whilst at sea. Hence, exchanging ballast water introduces a number of issues related to both stability and strength.

The advantages often mentioned are that no added or retrofitting of equipment is needed; it is a standard procedure for some vessels, the costs are relatively low for most vessels, and it is assumed fairly effective. Disadvantages often stressed are that weather conditions may effect safety, larger vessels mean higher costs, organisms in sediments are not removed, resistant life stages remain, and the method is unsuitable for near shore transport because of time limits and lack of deep ocean areas (see e.g. Woodward et al. 1994, AquaSense 1998). Other benefits and drawbacks are also addressed in several of the studies mentioned previously.

Exchange of ballast water is often referred to as the reference method for efficiency requirements for treatment options in general. It should be noted that the efficiency of ballast water exchange is not thoroughly verified. The method depends upon a number of factors and might alter as a function of ship-specific characteristics. Recent work (presented at the Conserted Action, "Testing Monitoring Systems for Risk Assessment of Harmful Introductions by Ships to European Waters", 6^th. workshop) clearly indicates that the method at this stage has several deficiencies when it comes to being a method of reference.

Hay & Tanis’ (1998) Cawthron report identifies three methods for exchanging ballast water: 1) re-ballasting, 2) overflowing of tanks, and 3) the Brazilian dilution method. An evaluation of these are listed in table 6.9.

Rigby & Hallegraeff (1994) did computer simulations which showed that re-ballasting caused bending moments or shear stresses that exceeded allowable values. Continuous flushing was considered both a safe and viable option on bulk carriers. Their testing of this procedure at sea on the bulk carrier "Iron Whyalla" found that after three tank volumes were exchanged, about 95 % of the original water was replaced. The estimated total exchange operation (with preparations) was approximately three days.

Armstrong et al. (1999) demonstrated that the use of computational fluid dynamics (CFD) was an effective design and analytical tool for finding a simple and cost-effective design method for exchange.

The IACS (1998) lists 15 hazards for re-ballasting and 12 hazards for flow-through exchange. These are summarised in table 6.10.

Woodward et al. (1994) investigated the safety aspects of exchanging ballast water in mid-ocean using three pilot cases; a container vessel, a tanker, and a dry-bulk carrier. Hydrostatic data were used to find still-water changes during exchange. The exchange scenario was then "exposed" to "at-sea" conditions. Hull bending moment, GM, and draft/trim were given particular attention. The study concluded that stability is not likely to be a problem, and that exchange can be done within acceptable safety margin, as long as wave heights are below a maximum value, suggested between 10 and 20 ft.

Occasionally the term "deep ocean" is used when talking about ballast water exchange at sea. In this context AquaSense (1998), and others, state that "deep ocean water contains species that will not easily adapt to coastal or fresh water environment" (p. 51). Indeed, as the IMO Guidelines recommend, ballast water is to be exchanged where depth exceeds 500 m, but it is important to note that water pumped into the vessel is that from the ocean’s surface layers and not from the deep. Whether organisms in the deep can adapt to other environments is thereby not an issue, unlike organisms from the surface layers of deep ocean areas’ which are. (The depths of mid-ocean areas can be fairly low (less than 500 m), and do therefore not fulfil the requirements of exchange.)

The literature study showed a great variety of possible options and combinations of options to treat ballast water. Each of the studies that compared methods had specified their own defined presuppositions, for example The Committee on Ships’ Ballast Operations et al. (1996) investigated on-board options only, and Laughton et al. (1992) only investigated The Great Lakes area. Some techniques were rejected for more or less unknown reasons (e.g. Laughton et al. 92). The treatment methods were evaluated by a number of criteria, and these were given different values between studies. For example, the Committee on Ships’ Ballast Operations et al. (1996) hardly considered costs, while others found this criterion to be important (e.g. Gollasch 1997, Laughton et al. 1992). There are also examples of studies focusing only on specific taxa (Bolch & Hallegraeff 1993, Rigby et al. 1993), and many rely only on theory (e.g. Laughton et al. 1992) while others test their theories in practice (Armstrong et al. 1999, Cangelosi 1997, Bolch & Hallegraeff 1993). In addition, there are works where one interesting technique is studied without comparing it with others (e.g. Lafferty & Kuris 1996). These different scenarios make comparisons difficult. In all, there seems to be an agreement that combinations of treatments are the most favourable solution, and in general, mechanical options were considered more favourable than chemical options.

Treatment of ballast water can be undertaken by several different methods either on shore or on board. Greenman et al. (1997), students exercise (Worcester Polytechnic Institute, UK) is mentioned here to illustrate necessary considerations considerations in relation to implementation of treatment facilities.

The work investigates treatment facilities as alternatives to mid-ocean exchange, and arrived at three options:

1. land-based facility as stopping-point for vessels
2. land-based facility with pipes or barges transporting ballast water to/from ships
3. floating facility.

Several criteria are listed for consideration when an optimal treatment method is to be chosen:

• land availability (if available: size, cost, ownership is to be evaluated)
• port layout (size, quantity of terminals, distribution of terminals)
• port traffic (number and types of vessels)
• volume deposited

When the above criteria suggest more than one solution, cost-efficiency and time-efficiency must be examined.

Cohen (1998), concludes in favour of on-shore treatment as apposed to a selection of on-board treatment options. Table 6.11 summarises the conclusions.

Filtration is mentioned positively by a number of relevant references. Filtration has been divided into different "sub-groups" in different studies. Carlton et al. (1995) distinguished between macro- and microfiltration, Gollasch (1997) considered self-cleaning filtration, microfiltration and granular filtration, and AquaSense (1998) added jet filtration.

Gollasch’ self-cleaning filtration is identical with the project on board the Algonorth (Cangelosi 1997) and has the advantages of being both effective as well as being safe for the ship, crew and the environment. The sediment problem is addressed by this method, and backwash is possible.

Microfiltration allows for a use of combinations of filters with different mesh sizes. Automatic self-cleaning is always a necessity because of rapid clogging and fouling of filters. It is predicted that pump capacity might be reduced because of such clogging (Gollasch 1997). Larger filter areas can reduce clogging as high flow rates are important during ballasting. A recirculation system for on-board treatment could also solve the problem (Laughton et al. 1992).

Jet filtration uses centrifugal force to separate sediment and organisms from the water (AquaSense 1998). Post-treatment with UV increases its efficiency.

Heat treatment introduces a long list of disadvantages, but is nonetheless considered positively in most of the studies (AquaSense 1998, Carlton et al. 1995, Committee on Ships’ Ballast Operations et al. 1996, Gollasch 1997), which seem mainly to concentrate on the method characteristics associated to safety, effectiveness for plankton, and operational costs.

The drawbacks mentioned are:

…necessary heat levels are not known; lower organisms may survive; even heat is difficult to achieve and heat loss is great; release of hot water might not be environmentally sound; impossible (in many cases) to adopt (cargo holds); may cause bacterial growth (sediment build up and bacterial corrosion); negative effects on pipes, pumps, and coatings; high costs of retrofitting; constraints on volume treated; length of voyage limits time for treatment; large energy requirements...

Many of these claimed disadvantages have been refuted (see Anonymous 1997, Bolch & Hallegraeff 1993, Hallegraeff et al. 1997, Rigby et al. 1997, 1998).

Traditionally, the method of exchanging ballast-water has been considered the only one possible to perform on most ships at relatively low costs. Now questions have been raised regarding the costs (Hay & Tanis 1998), and the de-ballasting and re-ballasting method has been challenged by the flow-through dilution exchange and the Brazilian dilution method. The latter two are claimed to be more efficient in removing organisms, safer, cost efficient, and with not much refitting of pipeworks required (AquaSense 1998, Armstrong et al. 1999, Dehalt 1999). However, this will probably depend more upon the actual vessel undergoing treatment.

The method of ozone is addressed separately as it is considered an exception to the group of chemical methods, not producing serious by-products which affecting the environment.

Oemcke & van Leeuwen (1998) investigated ozone for ballast water treatment, and reviewed different literature for comparisons. Here it was found that ozonation of seawater is different from that of ozonation of freshwater. Results of treatment of dinoflagellate hypnocysts were one of the objects of the study, but indicator organisms were used instead. Treatment in transit, during ballasting, and at shore-based facilities were evaluated.

In transit treatment:

• renders a longer treatment time possible
• is better operated as a completely mixed system
• a side-stream from the tanks is pumped through an ozone generator
• needs more than three recirculations
• may be at risk to corrosion and erosion
• power of about 20kW per kg of ozone required (relatively modest)
• costs are estimated to between US $ 2.5-9 million (depending on treatment time: 1-4 days)

During ballasting:

• all water will be exposed
• organisms that require short contact times are ensured disinfection
• the pumping rate of large bulk carriers is a big problem
• microbial activity in sediments will cause a locally ineffective disinfection
• corrosion will consume oxidant
• pre-filtration could be required
• costs are estimated to between US $ 1.5-10 million

Shore-based facilities:

• many practical problems
• possible to use large doses of ozone and pre-filtration
• long contact times still required
• possible to reduce costs by sizing plant and balance flows

Conclusions of Oemcke & van Leeuwen (98) were:

• results depend upon organisms and location for treatment
• treatment during ballasting or at shore-based plants is not useful for hypnocysts: the costs are too high, oxidant doses required are too large and corrosion will result
• to remove hypnocysts, filtration is recommended instead of ozonation during ballasting
• in transit treatment is effective for some dinoflagellates, amoebae, bacteria, and viruses, and could treat "tougher" organisms
• ozone could be used in combination with heat or filtration

From the methods considered, exchange filtration and heat treatment are techniques which have provided the most promising results. However, this may only reflect these methods being the most widely investigated. The availability of information on other methods has been more restricted. A vast amount of alternative options is therefore not considered very practical. This is often because they compromise safety (i.e. chemical treatment) or are not proven effective. Many of the conclusions may not be objective, for example those regarding costs. This calls for additional study of these techniques, as some, or even several of them, might "score higher" after additional investigations.

Many new techniques have recently emerged and some may prove to be better options than the existing ones. Further work on design, costs, safety, etc is required on both existing and new techniques, as well as practical trials on board vessels and in laboratories.

The formal safety assessment approach has until now been somewhat dominating in the discussions concerning the development of rules and regulations on the ballast water issue. The feasibility of replacing this with a risk-based approach has been debated by nations, but also within the IMO.

Based on a view that no species introductions should be allowed, many may favour a strict rule-based approach to regulate and control the use of ballast water. To achieve a successful rule-based approach will generate bureaucracy and require corresponding control mechanisms. The inability of rule-based systems to differentiate between cases where measures are needed to control risks and where they are not required, produces costs that benefit neither the environment nor shipping industry, governments or society in general. The high costs associated with the implementation of inefficient or unnecessary measures and strict control mechanisms are the main disadvantages of a pure rule-based system. Another problem is the time required by incorporating new knowledge into a rule-set. A case specific approach is more dynamic and can reduce costs, at the same time as the efficiency and reliability in dealing with ballast water issues are increased.

It is generally accepted that harmful aquatic transfers cannot be totally prevented anyway by regulations alone. The potential of improving the level of protection by evaluating planned ballast journeys with a risk-based approach should therefore be assessed.

This work proposes the use of a dynamic risk-based approach, which is flexible and allows case specific assessments. Simple cases can be treated quickly and generically. Potential "high risk" cases will be identified and subjected to detailed analysis. The collection of relevant case specific data provide for further studies and will include that of the efficiency of the implementation of various risk reducing measures. In this way, efforts can be concentrated on those cases which represent a greater risk. Relevant state of the art data and information from previous analysis are gathered, structured and stored in databases to provide experience feedback, and to maximise the efficiency and minimise the uncertainty in the risk assessments. The suggested approach is tailored to combine available historical data, including ballast water research with general and species specific biological knowledge, technical (e.g. vessel specific) and port/area specific knowledge and interdisciplinary expert judgements.

A detailed and logically structured process of identification and screening of potential hazards to identify all relevant risk contributors is a key feature of the risk-based ballast water management approach. Identification of the most critical risk contributors enables research, analysis and risk reducing efforts to be focused where the effect of utilising the resources can be maximised.

In 1995, the Australian Quarantine and Inspection Service (AQIS) established a national Ballast Water Management Strategy (Hayes and Hewitt 1998). This is the only other comprehensive work performed on the risk-based approach to ballast water transfer of harmful organisms. This section of the report compares EMBLA with the risk assessment framework developed to meet the objectives outlined in the Australian Ballast Water Management Strategy. The assessment of the Australian risk assessment framework is mainly based on information in Hayes and Hewitt (1998). Hayes (1997) was also consulted.

Generally, there are many similarities in the overall risk assessment approaches between the two methods. Neither EMBLA nor the Australian approach fully account for transfer of organisms on vessel exteriors (hull fouling).

The defined "endpoint" of the Australian risk assessment framework is "The introduction of a non-native organism into (an uncontaminated) port whose environment satisfies the bio-requirements of the organism at the time of introduction". This has the advantage of simplicity and implies that no assessment of the introduction being permanent or causing ecological or economical damage in the recipient port is necessarily required, as is the case in the EMBLA concept.

EMBLA and the Australian risk assessment framework both provides and represents:

• species and vessel specific analysis
• quantitative approach.
• utilise target lists to identify unwanted species.

Use of risk acceptance criteria is an integral part of the risk assessment framework. As considerations and interpretations of acceptable and unacceptable risk influence how a quantitative risk assessment is undertaken, risk acceptance criteria should be considered as early as possible in the risk assessment process. More work is required however to develop generally applicable and accepted acceptance criteria as standards built into the risk acceptance framework.

Acceptance criteria form the basis for every assessment as to whether a hazard or general scenario are potentially acceptable or not. The Australian approach "does not develop acceptance criteria, assess risk against these criteria nor consider treatment plans". Furthermore, the Australian framework "does not identify the existing management and technical systems to control risk". However, the report states that identifying procedures to control risk is also part of the framework objectives for the Australian work.

The Australian approach consists of a total of five levels aiming to provide an increasingly accurate assessment when moving from one level to the next. Although the approach does not define risk acceptance criteria, it states that it is not necessary to "progress to higher levels of assessment, if an acceptable level of risk is reached at an earlier level". This is in line with the EMBLA approach, where progress of the risk assessment is not required when the outcome is "acceptable risk". In total, the EMBLA concept extends further into the risk-based approach than previous work.

To ensure consistency in assessments, estimated risk must be compared with established and agreed risk acceptance criteria. As illustrated in this section, a wide range of methods to establish such criteria exist. The definition of "acceptable risk" is expected to vary between regions, for example due to national regulations (which are not always risk-based) and varying national perception, awareness and assessment of risk factors. Acceptance Criteria can be established based on acceptable probability or acceptable (insignificant or no) consequence alone. Risk-based acceptance criteria will usually be based on risk, i.e. the combination of probability and consequence.

Factors affecting risk acceptance criteria include:

• Local regulations and legislation in donor and/or recipient area, and whether these are respected or not.
• Ecological status of recipient area.
• Ecological status of the donor area, e.g. the presence of toxic algal blooms.
• Technical specifics of the vessel.
• Season of transfer.
• Perception of uncertainty.

The methodology is based on excluding events with an acceptably low risk from further detailed studies. This means that if an event is excluded from further study due to acceptable risk, the impact will be insignificant, regardless of the probability of the event occurring.

Examples of Risk Acceptance Criteria incorporated into the risk assessment of EMBLA are:

1. "If a toxic algae bloom is present in the donor port at the time of ballasting, the risk for introduction of alien species in the form of toxic algae to the recipient port will be considered unacceptable".
2. "Due to impaired environmental conditions in ports compared to the biogeographical regions they belong to, cases can exist where the donor and recipient ports are compatible even if the donor and recipient regions are not. Introduction of alien species not included in the recipient port’s target list (or hazard species list), or in the donor port’s hazard species list (called donor list) will therefore be possible. This is considered acceptable because such introductions, at this stage, are considered able of causing only a temporal local effect in the recipient port."
3. Introduction of non-indigenous species is acceptable to areas characterised as "evolution centres" and to regions with stable competitive ecology where the non-indigenous species are not capable of altering the local ecological balance.

In general, nations are expected to consider the introduction of any species on a relevant target list as unacceptable. This will then be considered unacceptable according to whether the case specific consequences of introduction of individual species on the target list are negligible or not.

Implementation of risk reducing measures will be considered when the potential risk is unacceptable. The efficiency of risk reducing measures is generally assumed to be vessel, journey and port specific. However, some risk reducing measures may also only be effective for some types of organisms due to their size, biological features and tolerance.

In the context of this report, a risk reducing measure is any measure that can prevent introduction or establishment of a hazardous species to the recipient area. The purpose of risk reducing measures are therefore to reduce an unacceptable risk level to an acceptable level. Risk reducing measures will normally be evaluated and implemented in the vessel during the operational phases involved:

1. Ballasting in donor port
2. Transfer in vessel
3. De-ballasting in recipient port.

In chapter 6, a comprehensive description and evaluation of existing proposed risk-reducing measures have been given.

Because some risk reducing measures are very expensive, the risk assessment approach decreases costs by effectuating the most appropriate and cost efficient case and vessel specific measures and avoids implementing risk-reducing measures when not required.

This section of the report details the risk assessment methodology developed to evaluate the risk of introduction of unwanted species connected to the transfer of ballast water in international shipping.

The present EMBLA risk assessment methodology is based upon four consecutive sections; the outcome of the first section determines whether the next section is required (a sequential approach). The four sections are:

1. Initial Hazard Screening
2. Detailed Hazard Screening
3. Hazard Analysis
4. Impact/Consequence Assessment

To ensure overall risk assessment methodology robustness, it is important to focus on the first section, the Initial Hazard Screening, and detail this to the highest degree possible, before the consecutive sections of the risk assessment are fully developed.

The framework upon which the risk assessment methodology and structure are based, consists of two parts; risk acceptance criteria and risk reducing measures. Upon completion of one step in the risk assessment methodology, a comparison between the result of that step of the assessment and risk acceptance criteria is used to determine whether further analysis is required. Where this is the case, a choice can be made to go to the next step in the risk assessment or to evaluate the efficiency of any relevant risk reducing measures.

The risk assessment methodology has an in-built flexibility for further development as more knowledge is gained (for example regarding the ecology of hazard species). The system utilises databases designed for storage and processing of the relevant data. The data structure and the GIS demo are described in separate chapters of this report.

Based on the current status of the risk assessment methodology, this section identifies and discusses current knowledge mismatch with respect to methodology and data availability. Priorities for future development of the database and methodology are suggested.

The risk assessment methodology framework as described in the previous report (DNV 1999) has been further developed, specified and integrated with the DNV GIS database system. The flexible structure of the database system is well suited for the adopted approach, with continuous updating as more information and knowledge is gathered.

The four main stages of the risk assessment methodology are:

1. Initial Hazard Screening (screening of biological factors)
2. Detailed Hazard Screening (including vessel/voyage and ballasting specific assessments)
3. Hazard Analysis (analysis of biological and case specific factors)
4. Impact/Consequence Assessment

The current state of development is described in the following sub-sections. A detailed description of the steps in the risk matrix is given in the appendix as a separate report with restricted distribution. Discussions concerning the current knowledge mismatch and need for further development are also presented.

The methodology is visualised in figure 7.1.

Figure 7.1 Illustration of the main sections of the risk assessment framework. Underlying acceptance criteria not included.

The automated Initial Hazard Screening in the GIS database system currently includes an assessment of the following factors:

Biogeographical compatibility

The biogeographical compatibility between the region of the ballasting port and the region of the de-ballasting port is assessed, based on the principles and classification of zoogeographical areas (zones, regions and provinces) as described in chapter 6 and the previous DNV report (DNV 1999).

• Compatibility between regions indicates a potentially unacceptable risk for the transfer of unwanted species.
• Incompatibility between areas indicates that the there is either no risk or acceptable risk for the transfer of unwanted species.

Based on biogeographical compatibility, transfer between two ports within one province, between two ports within the same region (when there are no defined provinces in the region), or from/to one single port, is considered to have an acceptably low risk level.

Historical Data:Hazard Species Lists and Target Lists

A comparison between the Hazard Species List for the region of the donor port and the defined Target List that covers the recipient port is performed. In cases where no relevant Target List has been produced, the assessment is undertaken by identifying a Hazard Species List for the recipient port.

The outcome of this comparison is called a Donor List. The Donor List is a list of potential hazard species that

• are known to be distributed in the region of the donor port, and
• are on the defined Target List (or if lacking a target list, on a Hazard Species List) covering the recipient port.

In cases where the recipient country/port do not accept introduction (or re-introduction) of any hazard species in general, all species on the Hazard Species List for the region of the donor port will automatically be included on the Donor List.

This is an assessment of whether a toxic algae bloom is present in the donor port or not. If that is the case, toxic algae will be added to the Donor List. At this stage of development of the risk assessment framework, presence of toxic algae in the donor port (which is the equivalent to ballasting location) is not considered acceptable. Risk reducing measures therefore have to be considered.

If neither of the three topics discussed:

1. Biogeographical compatibility,
2. Match of historical data for the donor and recipient regions and
3. Toxic algae bloom

are identified in the Initial Hazard Screening, the risk level will be considered acceptable. The donor and recipient ports are then considered incompatible. Further risk assessment will therefore not be required.

When the Initial Hazard Screening identifies potential unacceptable hazards (i.e. risk for transfer of hazardous organisms), the Initial Hazard Screening will also produce a Donor List, i.e. a list of potential hazard species. In this case, there are two alternatives:

1. Detailed Hazard Screening.
2. Evaluate implementation of Risk Reducing Measures, followed by a new Initial Hazard Screening. (Most relevant for toxic algae bloom.)

Below, major knowledge mismatches in the Initial hazard screening are listed.

The automated GIS-based Detailed Hazard Screening which is undertaken when the Initial Hazard Screening has identified a list of potential Donor Species and/or if the donor and the recipient areas are potentially "Compatible" currently includes an assessment of:

• Whether there is physical compatibility between the donor and the recipient port, with respect to species temperature and salinity survival requirements for any of the species on the Donor Species List. Requirements regarding port habitat are not included as general port habitat knowledge indicates that all usual habitat types will be represented in all ports.
• Whether there are freshwater species in the Donor Species List and a river outlet in the vicinity of the recipient port; an assessment of the physical compatibility between the temperature and salinity ranges of the river system and the identified freshwater species is then undertaken.
• Whether there are hard bottom species in the Donor Species List and an outlet system in the vicinity of the recipient port; an assessment of the physical compatibility between the temperature and salinity ranges of the outlet system and the identified hard bottom species is then undertaken.

The assessments take seasonal salinity and temperature variations into account as monthly averages. Compatible survival requirements related both to salinity and temperature are required for the port region, river outlets (if present in recipient port), or hard bottom species (assessment on outlet systems in the recipient port region) to enable these to survive.

One aim of the Detailed Hazard Screening is to identify case specific risk factors that will be evaluated in the risk assessment. These will be:

• The relevant case specific hazard species (Donor Species List). These species are considered capable of surviving in viable populations both in the donor and in the recipient ports during the time (month and month after) ballasting and de-ballasting occur.
• Environmental requirements; temperature and salinity regime and preferred habitat for the Hazard Species. The species will be split into groups according to their preferred habitat: port species; freshwater species (living close to or at river outlets) and hard bottom species (living in or close to outlet systems).

If the Donor List is not "empty" after the Detailed Hazard Screening, there are potentially unacceptable hazards (i.e. risk for transfer of harmful organisms). In such cases, there are two alternatives:

1. Undertake a Hazard Analysis, or
2. Evaluate implementation of Risk Reducing Measures, followed by a new hazard screening.

Below are listed major knowledge mismatches in the Detailed Hazard Screening.

The Hazard Analysis is a systematic and structured assessment of the relevant hazard species obtained from the Hazard Screening. The Hazard Analysis is not yet fully integrated into the GIS database. It would be advisable to keep it like this until the Initial and the Detailed Hazard Screening has undergone further development.

The Hazard Analysis is divided in three main sections according to the phases of use and transport of ballast water. Each section describes linkages in a chain through which the hazard species have to survive, to cause damage in the recipient area:

1. Ballasting
2. Transfer
3. De-ballasting

The approach is illustrated in figure 7.2 below. The main factors assumed to affect each part of the assessment are outlined in the sub-sections below. The likelihood of the species survival are assessed, based on species specific information (biology) and voyage/vessel specific information, together with donor and recipient port details, season of transfer etc.

The Hazard Analysis assesses the probability for survival of each defined hazard species through the three phases. The outcome of the Hazard Analysis is therefore the probability for survival of each of the defined hazard organisms. If the probability for survival of the defined hazard organisms is below the relevant acceptance criteria, the species can be excluded from further assessment. If the probability for one or more hazard organisms to survive the three phases of ballasting, transfer and de-ballasting, or the aggregate probability for all hazard species is potentially unacceptable, there are two alternatives:

1. Undertake an Impact/Consequence Assessment, or
2. Evaluate implementation of Risk Reducing Measures, followed by a new hazard analysis.

This part of the Hazard Analysis evaluates the probability of the defined potential hazard organisms to the following items:

1. Availability in the water column at the specific time and location of loading.
2. Survive the ballasting process and enter the ballast tanks in viable numbers.

This include assessments of various factors in the ballasting sequences where the following can be mentioned:

• Sunlight, current tidal range and ballasting depth, sea depth at ballasting location (assumed to affect 1. item above)
• Mesh size (characteristics) of ballast water intake and vessel specific characteristics of the ballasting procedures (ballasting pipeline system, suction pressures and velocities) (assumed to affect item 2 above).
• Environmental conditions in the donor port (fauna characteristics, ecological competitive conditions, sediment type/substrate, local habitat and time specific temperature and salinity ranges, proximity to sewage outlet, season, traffic in port etc.).

The biology (e.g. life cycle stage, vertical migration patterns, seasonal migration etc.) for each defined hazard organism will be evaluated with respect to technical conditions concerning the ballasting sequence steps such as:

1. Depth of intake of ballast water.
2. Time of day.
3. Filtration systems.

So far only step 1 and 2 have been incorporated into the GIS system.

Transfer Phase

The transfer phase of the Hazard Analysis evaluates the probability for potential hazard organisms to survive in the ballast tank during the voyage.

The following factors are evaluated with respect to vessel and voyage specific parameters determining the conditions in the ballast tanks during transfer:

1. Temperature and temperature variations in ballast tanks. This might be a function of vessel type, location of ballast tanks in the vessel (sea temperature exposure, proximity to warm areas such as engine rooms, hull type (single/double).
2. Degree of motion, vessel movements cause mixing of ballast water
3. Light, salinity, nutrient and oxygen content of ballast tanks
4. Duration of voyage.

Factors affecting the survival of organisms are assumed to include but not be limited to the following items:

• Temperature tolerance
• Darkness tolerance
• Nutrition requirements
• Reproductive biology/duration of life cycle stages etc.
• Oxygen level

De-ballasting phase

The de-ballasting part of the hazard analysis evaluates the probability for the defined hazard species to:

1. Survive the de-ballasting process and exit the ballast tanks.
2. Establish in the recipient area in viable numbers.

Factors affecting the survival and establishment of each defined hazard organism include:

• Mesh size (characteristics) and location of ballast water outlet in the various ballast tanks (affects probability for sediment and possible cysts to exit the ballast tanks) together with vessel specific characteristics of the de-ballasting procedure (pipeline system, pressures and velocities).
• Species’ reproductive strategy, life cycle stage at arrival and their tolerance to chemical and physical conditions in the recipient port.

These factors will be evaluated with respect to relevant environmental conditions for de-ballasting:

1. Location of de-ballasting in port.
2. Environmental conditions in port (local seabed conditions and fauna characteristics, sea depth, local and time-specific temperature and salinity ranges, proximity to sewage outlet, season etc.).

Both item 1 and 2 have been incorporated into the GIS system.

Hazard Analysis - Knowledge Mismatch

Below are major knowledge mismatches in the hazard analysis are listed.

Risk is a function of probability the of an event occurring and the potential consequences of the event. The impact/consequence assessment is the defined endpoint of the risk assessment, and has been separated into three main categories:

1. Ecological Impact
2. Economic Impact
3. Vessel Safety

The introduction of a hazardous species may cause an impact on one or more of the three defined impact categories. Impacts are generally divided into primary and higher order impacts. Primary (or direct) impacts occur as a direct consequence of the activity (for example a de-ballasting operation), while higher order (or indirect) impacts result from changes in a chain of environmental parameters (Wathern 1988). Impacts can also be cumulative, for example giving a larger effect over time than anticipated for direct impacts (e.g. dose response). The classification of impacts in categories depend on factors such as; the biology of the hazard species in consideration, the sensitivity of the recipient port and surrounding areas, and local legislation. Impacts will be assessed with respect to severity and duration, based on experiences from previous introductions.

Ecological Impact

Ecological disturbances, for example changes in the pre-existing biota and loss in biodiversity are examples of relevant ecological impacts.

The severity of ecological impacts, which can be a function of for example the speed and duration of an invasion, is separated into three categories:

• High ecological impact. This implies an invasive spread of the hazard species and the uncontrolled spread of this organism, leading to loss of ecological diversity (other species e.g. keystone species becoming extinct). Introduction of the considered hazard species may cause secondary or cumulative impacts (for example further geographical spread of the species).
• Medium ecological impact. The hazard organism has a high rate of spread, but it does not have the potential to dominate the local ecosystem and will therefore have a limited effect on the ecosystem balance.
• Low ecological impact. The hazard organism does not have the potential to dominate the local ecosystem. It can co-exist in balance with the local ecosystem without affecting the dynamic structure and balance of the system to any noticeable degree. Low severity will frequently be defined as acceptable from an ecological viewpoint, but this will vary with local acceptance criteria. If the recipient area is a Marine Protected Area (MPA) or a Special Area, no permanent introduction of any alien species will be acceptable.

The duration of the impact will be separated in two categories:

1. Seasonal. Non-permanent impacts, i.e. the ecosystem damage can be reversible.
2. Permanent. For simplicity, permanent damage has been defined as damage that will last at least one year and which is non-reversible.

Economical Impact

Economic loss, for example for fish farms and commercial fisheries, or problems associated with the fouling of tubes and pipelines, caused by the introduction of alien species, are examples of possible economical impacts. The severity of economical impacts will be separated in three general categories:

• High economical impact. Permanent and serious economical damage of existing industries and their economic basis. One example is an impact which is detrimental for fish farming or local fisheries, another the clogging of industrial outlets.
• Medium economical impact. Limited permanent damage, limited damage of limited or long duration or more serious damage of seasonal (short) character.
• Low economical impact. Economical impact that is considered acceptable based on local acceptance criteria.

The duration of the impact will be separated in two categories:

1. Seasonal/Immediate. Non-permanent impacts i.e. the economical damage can be recovered.
2. Permanent. Non-reversible impacts. A permanent impact does not have to be immediate

There is no automatic link between high ecological and high economical impacts. A seasonal ecological damage can for example cause permanent economical damage on existing facilities (and be detrimental to for example fish farming).

Vessel Safety

The third impact category - Vessel Safety may need to be considered, for example when the assessment includes risk reducing measures such as re-ballasting during transfer or other measures that might influence the operational safety aspects of a vessel. The vessel safety issue involves both technical safety of vessel and safety of the crew.

Knowledge Mismatch – Impact/Consequence Assessment

Listed below are major knowledge mismatches in the Impact / Consequence assessment listed.

This section summarises the knowledge mismatch related to the risk assessment and recommends priorities for further work.

As emphasised in previous sub-chapters, the main limitations of the risk assessment are currently associated with lack of data. No historical database enabling risk quantification of ballast water management strategies has to date been gathered. As assessments have to be made, based on expert judgements and available data/experience, gathering, building and verifying historical databases are important factors, both of ongoing, and future work.

Contents of the current databases need further testing, verification and update as more information is made available. It is recommended that the risk assessment methodology is thoroughly tested by verification runs against known and suspected transfers.

At present time, the system deals with risk assessment of single transfers. However, as discussed the risk assessment structure is suitable for the assessment of multiple journeys.

As discussed in greater detail, species ecology is neither fully known nor understood. Furthermore, the interactions between different species (e.g. populations’ growth and decline) and between species and their physical environment (e.g. stress, salinity and temperature tolerances) are numerous and complex. Therefore, not even a comprehensive historical database can ever be expected to be complete. A process of continuous updating and improvement is therefore needed.

Currently, the risk assessment matrix focus on transfer of organisms via ballast water in ballast tanks. Carlton et al. (1995) identified a total of nineteen separate transport vectors associated with shipping. The transfer of organisms on ship hulls is the second of these transport vectors considered to be of major importance. This transport vector has not yet been specifically included in the risk assessment. Hull-species included in the hazard species list are expected to be identified in the current hazard identification process. Extending the risk assessment to include hull species is therefore not expected to require altering the hazard screening. Different factors will need to be assessed in the hazard analysis as the ballasting and de-ballasting processes in themselves become irrelevant.

The methodology for the estimation of probabilities of unwanted events, categorisation of consequences, and definition of acceptance criteria must be continuously refined as more knowledge is gained.

Challenges include:

• How to extract and systemise reliable biological information from available sources (state of the art), and how to accept and quantify uncertainty. Problems include methodological inconsistency between data sources (e.g. different sampling methods, lack of replicates to account or quantify biological variation versus sampling errors), and total lack of available data sources.
• Quantification of risk and uncertainty.
• How to deal consistently with cases where data is seriously lacking. At present, the development of a standard set of conservative "base-case" values is recommended.

A quantitative approach has been adopted to evaluate risk in the form of probabilities for hazard species to transfer to and establish in new environments (Hazard Analysis) and the potential consequences of such establishment (Impact/Consequence Assessment), but more work is needed to establish the required methodology.

An approach without quantification will be more "open" to subjective interpretations and value judgements and more difficult to standardise the methodology required for a successful risk-based database system. Due to the high number of variables that might influence the total risk and the expected geographical acceptance criteria variations, comparison of different cases will be more difficult with a qualitative approach. Adopting a quantitative approach will make it easier to compare the risk associated with for example different trade routes, harbours or countries.

One aim of quantitative methods is to express all relevant impact in a form that can be aggregated. Numerical results offer a good basis for clear result presentation, and have the potential to overcome subjective risk-perception problems. One example is that low probability/high consequence events generally tend to be overestimated, at the same time as high probability/low consequence events tend to be underestimated. Table 7.1 gives an example of a possible categorisation of probability ranges for the risk assessment.

To use of and interpret probability ranges in the hazard analysis, it is important to note that the probability for one defined hazard species (species i) to be assessed as potentially unacceptable can be expressed as:

P_{species i Hazard} = P_{species i Ballasting} * P_{species i Transfer} * P _{species i De-ballasting}

To express the total hazard, the aggregated probability of all hazard species being unacceptable have to be considered, i.e.:

P_total _{all Hazard species} = å (P_{species i hazard}) _{all Hazard species}

The outcome "Potentially unacceptable" from the Hazard Analysis can therefore indicate high contribution from one hazard species only, a relatively small contribution from a high number of hazard species or a combination of the two.

The following discuses the availability of data and sources.

Historical Data

Previous findings for example, historical records of transfers or preferably a database of historical records is the most important sources of information required for a successful risk assessment. Current historical data and target lists therefore need frequent reviewing. It is essential that the ecological information regarding the various hazard and target species are kept as complete and accurate as possible. A system for the continuous updating of historical information (as more is made available worldwide) has to be implemented into the computerised DNV GIS system.

Further research into the ecology of species recorded on the existing historical and target lists will therefore be required to reduce the current knowledge mismatch for all parts of the risk assessment. Gathering and compilation of historical data including critical reviews of source information (how likely is it that the species transfer actually occurred via ballast water?) should be performed.

Research will have to include sampling of ballast water and experimental work. Development of standardised methodology for ballast water sampling will enable comparison of data from different sources, and ensure that the quality of the sampling exercises is that which provides reliable information.

Compatibility and Biogeographical Regions (GIS Database)

Together with the historical data, the definition and use of biogeographical zones, regions, provinces and the compatibility between these, represent important inputs to the Initial Hazard screening.

To what extent further refinement/expansion and/or verification of the biogeographical areas would:

1. be possible, and
2. result in more reliable results

for the risk assessment should be evaluated. This would include the benefit of refining the compatibility assessment in the GIS database screening to include areas defined as particularly vulnerable (for example a higher number of provinces). The same result would possibly be obtained by refining the use of temperature and salinity ranges.

There are several separate options or a combination of options that might be feasible:

1. Verification of the "current" biogeographical compatibility, including the borders defining the geographical boundaries between zones, provinces and regions. The biogeographical compatibility is based on benthic fauna. Refinement of the approach, for example, based on research findings might be appropriate. Compatibility assessments should be verified by a third party.
2. Expand the assessment of biogeograpical compatibility to include defined vulnerable areas.
3. Develop further the criteria for compatibility between ports to improve the port compatibility assessments in the Detailed Hazard Screening, including an assessment of which port- specific data that should be incorporated in the GIS port database.
4. Use of historical data and target lists in the Initial Hazard Screening. This would include further development and verification of how historical information is linked with port/region/province.

The database system developed for the EMBLA concept is based on the system used in the Pre-study; MS Access and the GIS of ArcView. The final product is expected to be converted to a more powerful and heavier database tool e.g. SQL server or Oracle. However, for the purpose of demonstration and testing, the adopted software performs satisfactory.

The data structure is flexible, allowing modifications according to future needs and specifications. The structure of the database is relational, with information stored in several tables, which are appropriately related.

The major tables are described below. The complete data structure is presented in the appendix as a separate report, with restricted distribution.

To perform the compatibility assessment, information related to biogeography is used. The establishment of biogeographical divisions is based on literature study and knowledge related to biological and oceanographic processes. The established biogeographical divisions and links between zones, regions and provinces are given.

The Hazard Species List and Target List are lists of species and species related information (e.g. temperature and salinity tolerance, habitat preferences, known ecological and economical impact). The lists will be used for the purpose of identifying compatibility between donor and recipient ports of ballast water. The type of information in the hazard species list and target list is identical, however this may differ with respect to the type of species listed. The information reflects parameters necessary for the purpose of performing the compatibility assessment. The total contents of lists and sub-tables thereof are given in the appendix.

Information on ports is used in Initial Screening, Detailed Screening and Detailed Hazard Assessment. The port table contains information such as temperature and salinity regimes in the port on a monthly basis, river outlets in the port region, and links to biogeographical divisions in the Hazard Species List. Only a selected number of ports are included in the Port table for demonstration purposes.

The information reflects the parameters necessary for performing the risk assessment. The total contents of lists and sub-tables thereof are given in appendix.

Information on vessel level is required to perform the Hazard Analysis. The type of information requested is for example departure month at Donor Port; arrival month at Recipient port; ballast water volume discharged; number of ballast tanks; voyage duration; vessel type; name; IMO-number and ship owner.

The log-database will contain results from all performed assessments of the EMBLA system. All information given for each case will be stored for statistical presentation purposes. Equally, all results from the risk assessment will be stored. This information can be used for control and verification of previous assessments and the presentation of statistics of ballast water transfer. In many cases a vessel performs the same journey over and over again. Where this is the case the vessel can refer to previous assessments and conclusions thereof. If there are any changes in information related to the vessel, voyage, ecological features of the ports (donor and recipient) or organisms in the Hazard Species List or Target Lists a new risk assessment should then be performed.

The demonstration kit developed as part of the Pre-study has been further detailed.

The kit is developed in the database system of MS Access and the geographical information system (GIS) of ArcView, accessing a set of databases and tables. Although the main functionality of ArcView as an analysing tool has not yet been applied, the software system has a very good user interface. Some of the information input given to the risk assessment is geographically linked; the system of ArcView is in such cases ideal for presentation purposes. It is also expected that in the future the analysis tools available in ArcView will be applied/ modified. Presently, the risk assessment procedures are undertaken using MS Access, leaving ArcView used only for presentation purposes. The GIS part of the system is the same as that developed under the Pre-study. The Access part however, has been further developed during the Integration phase.

The EMBLA demonstration kit is established and based on a selected case of transfer of ballast water. The vessel M/V Taronga was used for the purpose. Barber Ship Management and Wilh. Wilhelmsen have kindly contributed with important information for the production of the demonstration kit.

The M/S Taronga is a 40 000 DWT RoRo vessel and follows a regular trade pattern from Europe to Japan and back, via USA and Australia, making a total of 35 calls. The time schedule of the M/S Taronga during this voyage is 5 months (June – October).

The trade pattern of M/V Taronga was provided and a number of eight ports selected for the purpose of the demonstration kit. The following ports have been included in the demonstration kit and risk assessments have been performed for travels between these ports.

1. Port of Rotterdam, Netherlands
2. Savannah, USA
3. Auckland, New Zealand
4. Brisbane, Australia
5. Fremantle Australia
6. Kobe, Japan
7. Los Angeles, USA
8. New York, USA

The criteria for selection are based on availability of information at an acceptable level of detail. In addition, the selection is based on presenting an existing trade where the issue of ballasting is relevant. Focus on the ballast water issue at the relevant donor and recipient countries are also high. For each port, relevant information such as salinity and temperature gradients to perform the risk assessment is collected. Fairplay Publications has kindly retrieved information from their database system. This information has been loaded into the port database of EMBLA.

The results of the risk assessments for the following stages in the route of M/V Taronga are presented below.

1. Savannah, USA - Port of Rotterdam, Netherlands in October
2. Savannah, USA – Auckland, New Zealand with departure in August and arrival in September

It should be noted that the risk assessment of the EMBLA is not fully developed and the cases given below are merely to present the principles of EMBLA. The EMBLA concept can presently not give reliable risk assessments of ballast water transfer.

On its return to Europe the M/V Taronga leaves Savannah, USA in September and arrives at Port of Rotterdam in October. During this voyage the vessel makes three other calls in Europe prior to Port of Rotterdam.

The results of the EMBLA risk assessments for transfer of ballast water from Savannah to Port of Rotterdam is given below.

The Initial Hazard Screening indicates that the risk is not accepted since species in the Donor Port (Donor List) are listed in the Target List of the Recipient Port as indicated under the heading "Species in both lists". Although the Recipient Port has addressed that it cannot accepted to receive these species it could be relevant to identify the risk-level of successful transfer. The next step of Detailed Hazard Screening is performed.

In the Detailed Hazard Screening, temperature and salinity gradients of the Recipient Port is compared with temperature and salinity regimes of the species identified in Donor List. Since the gradients overlap the risk is unaccepted. The next step of Hazard Analysis is performed.

In the Hazard Analysis the species’ success in entering the vessel, surviving (High, Medium or Low) the different steps in the transfer process (a total of 4 steps) is assessed. This information should be available in the Target List and Hazard Species list. If that is not the case, the figures can be given manually. In addition, the level of eutrophication in the Recipient Port is implemented. As illustrated, the risk is given as unaccepted due to high probability for survival. The next step of Impact/Consequence Assessment is performed.

In the impact/consequence assessment, issues related to ecological, economical and easthetic impacts are assessed. Information on volume of ballast water discharged, rate of spread of organisms, temperature and salinity regime in surrounding areas of the port are among other relevant parameters to be assessed. In addition, issues related to vessel safety will be implemented in the Impact assessment phase.

The journey from Savannah to Port of Rotterdam has an unacceptable risk with respect to transfer of organisms in ballast water unless risk-reducing measures are undertaken.

The step of risk reducing measures is yet to be discussed and defined in the EMBLA concept.

M/V Taronga leaves Savannah, USA, passes through the Panama Canal and arrives in Auckland, New Zealand in July.

The results of the EMBLA risk assessment are given below.

As can be seen the result is indicated as accepted in the first step of Initial Hazard Screening. Here the biogeographical regions of donor port and recipient port are screened. In addition, a comparison between the Donor List (harmful species identified in the donor port region) and the Target List (list of unwanted species in the Recipient Port) is performed. The results of the latter are given under the heading "Species in both lists". As can be seen this list is empty. The transfer is therefore accepted, although harmful species are identified in the Donor Port region as indicated under the heading "Species identified in the donor region and not in recipient region". In this case the Target Lists of the Recipient Port should be updated. The system will therefore not only be able to assess risk but also be helpful in establishing Target Lists.

The findings presented in the previous pre-study and further assessed in this work have provided a basis for addressing the issue of developing the EMBLA concept. Development is planned throughout three stages (EMBLA 2000/ 2001/ 2002). The development stage EMBLA 2000 will be focused upon here.

During the progress of this work, main areas requiring particular attention has been identified:

• Infrastructure and user interface (IUI)
• Data processing and analysis (DPA)
• Standards and norms (SAN)
• Consequence assessment methodologies (SAM)

The following will briefly describe identified challenges associated to these and provide a "short term" priority as to development sequence. Note that the areas or tasks above have been addressed in project memos and in correspondence to partners.

EMBLA will provide quarantine clearance services to an international moving industry operating on flexible schedules. The nature of the industry will put requirements to the system in relation to its ability to respond. The system will require input from the vessel and must enable communication to port states or to other relevant authorities.

Provision of the EMBLA service will depend on use of the Internet. Evaluations incorporated in the concept are highly automated and performed electronically. A realisation of EMBLA will attempt to utilise standard available tools. These might require adjustments or modifications. However, the specifics of the requirements related to infrastructure and user interface can not at this stage be identified so development efforts on this topic are therefore at present not feasible.

The IUI topic requires input from the EMBLA’s actual procedure requirements in relation to:

• Data input and system updating
• Risk assessment procedures including screening and analysis stages
• Consequence analysis stages
• Output communication

The EMBLA Integration phase has developed a stepwise risk matrix assessment approach (SRA). This approach identifies the data input requirement associated to parameters and data sources (donor port; voyage; recipient port; target lists; historical data; etc.). Further, it addresses relationships between these and identifies driving parameters as a function of scenario specifics (season, region, etc.).

The EMBLA SRA at this stage presents an assumed invasion mechanism, where required presence of various factors are in some detail included. This mechanism is not thoroughly investigated.

A detailed evaluation of the SRA stages will be required. Laboratory activities and the establishment of data collection routines will be necessary.

The integration phase has also identified required data structure flow (DSF) supporting the various assessment stages involved. This connects the various assessment procedures, as well as organising and allocating necessary data. The generation of reporting formats and assessment logs are included. The data flow requires system functionality testing.

The following development stage of EMBLA (EMBLA 2000) will address both SRA and the DSF. A work programme incorporating the two will be based upon the findings of this report and its annexes.

Work has revealed an almost total lack of mutal-agreed procedures in this area. Three topics are of particular concern.:

• Standards associated with the sampling of ballast water (EU Concerted action)
• Standards concerning the reporting of invaders/ identifying particular organism of concern (Target lists/ Hazard species lists)
• Standards with reference to requirements on risk reducing measures. Exchange of ballast water has been suggested (by IMO) as a reference standard. Unclarities associated with this method as well as with the feasibility of using such a reference need to be addressed/ dealt with.

EMBLA 2000 will investigate procedures to arrive at standards on these items. Priority will be given to the latter (standards with reference to alternative measures). This is directly linked to the module of risk reducing measures in the SRA.

Three factors forming "total consequence" have been identified. These are all linked to "acceptance criteria’s" and need to be further defined.:

• Ecological consequence
• Economic consequence
• Safety

The forthcoming project stage will define the scope of all three steps. The latter, concerning safety will be including standards associated to treatment options. Safety interference related to the introduction of risk reducing measures will be addressed in a separate work supporting EMBLA 2000. This will firstly assess limitations related to the vessel and operational procedures linked to treatment requirements. Risks associated to the exchange of ballast water will be addressed in particular.

Anon. 1997. Heat treatment solution for infected ballast water. MER-Marine Engineers Review, p. 15.

Armstrong, G, Rose, A & Holdø, A 1999. An analysis of flow-through ballast water exchange. MER-Marine Engineers Review, pp. 45-46.

AQIS (Australian Quarantine and Inspection Service) 1993. Ballast water treatment for the removal of marine organisms. Report No.1. Ballast Water Research Series. Canberra, Australia.

AquaSense 1998. Risk assessment ballast water. Ecological and economic consequences, frequency analysis and measures. Ordered by the Dutch Ministry of Transport, Public Works and Water Management, North Sea Directorate. No 98.1290, 71 pp.

Bolch, CJ & Hallegraeff, GM 1993. Chemical and physical treatment options to kill toxic dinoflagellate cysts in ships’ ballast water. J. Marine Env. Engg. 1, pp. 23-29.

Briggs, J. 1974. Marine zoogeography. University of South Florida.

Cangelosi, A 1997. Battling biological pollution. 7 pp.

http://www.nemw.org/ERballast.htm

Carlton, JT 1985. Transoceanic and interoceanic dispersal of coastal marine organisms: The biology of ballast water. Oceanogr. Mar. Biol. Ann. Rev. 23, pp. 313-371.

Carlton, JT & Geller, JB 1993. Ecological roulette: The global transport of nonindigenous marine organisms. Science 261, pp. 78-82.

Carlton, J. 1993: Biological invasions and biodiversity in the sea: The ecological and human impacts of non-indigenous marine and estuarine organisms. In Proceedings of the conference & workshop: Nonindigenous estuarine and marine organisms (NEMO). US Department of Commerce. Seattle, Washington.

Carlton, J.T., Reid, D.M. and van Leeuwen, H. 1995. The role of shipping in the Introduction of Nonindigenous Aquatic Organisms to the Coastal Waters of the United States (other than the Great Lakes) and an Analysis of Control Options. Report No. CG-D-11-95, National Technical Information Service, Springfield, Virginia 22161.

Carlton, JT, Reid, DM & van Leeuwen, H 1995. Shipping study. The role of shipping in the introduction of nonindigenous aquatic organisms to the coastal waters of the United States (other than the Great Lakes) and an analysis of control options. The National Sea Grant College Program/Connecticut Sea Grant Project R/ES-6. Department of Transportation, United States Coast Guard, Washington, D.C. and Groton, Connecticut. CG-D-11-5, Government Accession Number AD-A294809. 213 pp. + Appendices A-I (122 pp.)

Chesapeake Bay Commission 1995. The introduction of nonindigenous species to the Chesapeake Bay via ballast water. Strategies to decrease the risks of future introductions through ballast water management. 31 pp.

http://nas.er.usgs.gov/publications/ballast.htm

Chu, KH, Tam, PF, Fung, CH & Chen, QC 1997. A biological survey of ballast water in container ships entering Hong Kong. Hydrobiologia 352 (1), pp. 201-206.

Cohen, AN 1998. Ships' ballast water and the introduction of exotic organisms into the San Francisco Estuary: Current status of the problem and options for management. A report for the CALFED Category III Steering Committee administered by the California Urban Water Agencies. Report No. 81.

Committee on Ships' Ballast Operations, Marine Board, Commission on Engineering and Technical Systems, & National Research Council 1996. Stemming the tide: controlling introductions of nonindigenous species by ships' ballast water. National Academy Press, Washington, DC, 141 pp. ISBN 0 309 05537 7.

DNV (1999). Ballast Water Transfer Atlas, Hazard Assessment and Decision Support Pre-study. DNV Report 99-3166 for Norwegian Research Council and MARMIL.

Dehalt, A 1999. The Brazilian dilution method. Ballast Exchange, volume 1, fall 1999, p. 9. West Coast Ballast Outreach Project.

DEL Industries 1997. Application Guidelines, Commercial Pools & Spa Pools, Corona Discharge Ozone Systems, No 3, 22 pp.

Delta Marine International, Inc. 1999.

http://www.chem-freeozone.com

Designers & Planners, Inc. & Herbert Engineering Corp. 1998. An investigation into the impacts of performing a ballast water exchange at sea on typical PANAMAX containerships. Contract: DTCG23-95-D-HMT001, D.D: DTCG23-95-F-HMT001, 11 pp.

Dickman, M & Zhang, F 1999. Mid-ocean exchange of container vessel ballast water. 2:Effects of vessel type in the transport of diatoms and dinoflagellates from Manzanillo, Mexico, to Hong Kong, China. Marine Ecology Progress Series 176, pp. 253-262.

Dodgshun, T & Handley, S 1997. Sampling ships' ballast water. A practical manual. Cawthron Report 418, 46 pp.

Ekman, S. 1953. Zoogeography of the sea. Sidgwick and Jackson Limited London.

Galil, BS & Hülsmann, N 1997. Protist transport via ballast water - Biological classification of ballast tanks by food web interactions. European Journal of Protistology 33 (3), pp. 244-253.

Gollasch, S 1997. Removal of barriers to the effective implementation of ballast water control and management measures in developing countries. IMO report, 195 pp.

Greenman, D, Mullen, K & Parmar, S 1997. Ballast water treatment systems: A feasibility study. U.S. Coast Guard & Worcester Polytechnic Institute, 59 pp. http://www.imo.org/imo/links/lnkstart.htm

Gollasch, S, Dammer, M, Lenz, J & Andres, HG 1995. Non-indigenous organisms introduced via ships into German waters. International Council for the Exploration of the Sea ICES CM 1995/0, 21 pp.

Gollasch, S 1996. Untersuchungen des Arteintrages durch den internationalen Schiffsverkehr unter besonderer Berücksichtigung nichtheimischer Arten. Verlag Dr. Kovac, Hamburg (Diss.), 314 pp.

Gollasch’ homepage: http://members.aol.com/sgollasch/sgollasch/index.htm

Gollasch, S, Lenz, J, Andres, HG & Dammer, M 1996. Introduction of non-indigenous organisms into the North and Baltic Sea: Investigations on the potential ecological impact by international shipping. 3 pp.

(http://www.ifm.uni-kiel.de/pl/transpor/default.htm)

Gollasch, S 1997. Removal of barriers to the effective implementation of ballast water control and management measures in developing countries. IMO Report, 195 pp.

Gollasch, S, Rosenthal, H, Laing, I, Leppäkoski, E, Macdonald, E, Minchin, D, Nauke, M, Olenin, S, Utting, S, Voigt, M & Wallentinus, I. 1998. Survival rates of species in ballast water during international voyages. Results of the first intercalibration workshop on ship sampling techniques and the first ocean-going workshops of the European Concerted Action: "Testing Monitoring Systems for Risk Assessment of Harmful Introductions by Ships to European Waters". 21 pp.

Gollasch, S 1999. Summary of the fourth land-based workshop of the EU Concerted Action "Testing Monitoring Systems for Risk Assessment of Harmful Introductions by Ships to European Waters" Conwy, Wales. 58 pp.

Hallegraeff, GM & Bolch, CJ 1991. Transport of toxic dinoflagellate cysts via ships ballast water. Marine Pollution Bulletin 22 (1), pp. 27-30.

Hallegraeff, GM & Bolch, CJ 1992. Transport of diatom and dinoflagellate resting spores in ships' ballast water: implications for plankton biogeography and aquaculture. Journal of Plankton Research 14 (8), pp. 1067-1084.

Hallegraeff, GM, Valentine, JP, Marshall, J-A & Bolch, CJ 1997. Temperature tolerances of toxic dinoflagellate cysts; application to the treatment of ships’ ballast water. Aquatic Ecology 31, pp. 47-52.

Hallegraeff, GM, Bolch, CJ, Koerbin, B & Bryan, J 1998. Ballast water a danger to aquaculture. Australian Fisheries 47 (7), pp. 32-34.

Harvey, M, Gilbert, M, Gauthier, D & Reid, DM 1999. A preliminary assessment of risks for the ballast water-mediated introduction of nonindigenous marine organisms in the Estuary and Gulf of St. Lawrence. Canadian Technical Report of Fisheries and Aquatic Sciences 2268, 56 pp.

Hay, C, Handley, S, Dodgshun, T, Taylor, M & Gibbs, W 1997. Cawthron’s Ballast Water Research Programme Final Report 1996-97. Cawthron Report No. 417, 144 pp.

Hay, CH & Tanis, D 1998. Mid-ocean ballast water exchange: Procedures, effectiveness, and verification. Cawthron Report 468, 49 pp.

http://www.cawthron.org.nz/downloads.htm

Hayes, K.R. (1997). A review of ecological risk assessment methodologies. Technical Report Number 13. Centre for Research on Introduced Marine Pests.

Hayes, K.R. and Hewitt, C.L. (1998). Risk Assessment Framework for Ballast Water Introductions. Technical Report Number 14. Centre for Research on Introduced Marine Pests.

Hines, AH & Ruiz, GM 1998. Biological invasions of cold-water coastal ecosystems: ballast-mediated introductions in Port Valdez/Prince William Sound, Alaska. Smithsonian Environmental Research Center Progress Report 26, 37 pp + appendices.

IACS (International Association of Classification Societies) 1998(?). Preliminary hazard identification. Ballast water exchange at sea. Annex.

ICES 1999. Report of the ICES/IOC/IMO study group on ballast water and sediments, Llandudno, UK 12-13 April 1999. ICES CM 1999/ACME: 2 Ref.: E, 22 pp.

IMO 1997. Guidelines for the control and management of ships’ ballast water to minimize the transfer of harmful aquatic organisms and pathogens. IMO Assembly Resolution 868 (20).

Ichikawa, S, Wakao, Y & Fukuyo, Y 1992. Extermination efficacy of hydrogen-peroxide against cysts of red-tide and toxic dinoflagellates, and its adaptability to ballast water of cargo ships. Nippon Suisan Gakkaishi 58 (12), pp. 2229-2233. (in Japanese with English summary)

Kabler, LV 1996. Ballast water invaders: Breaches in the bulwark. Aquatic Nuisance Species Digest 1 (3), pp. 33-35.

Kelleher, G. & Kenchington, R. 1992: Guidelines for establishing marine protected areas. A marine conservation and development report. Gland, Switzerland: IUCN.

Kelleher, G., Bleakley, C. & Wells, S. 1995: A global representative system of marine protected areas. The Great Barrier Reef Marine Park Authority, the World bank and The World Conservation Union (IUCN).

Kelly, JM 1993. Ballast water and sediments as mechanisms for unwanted species introductions into Washington State. Journal of Shellfish Research 12 (2), pp. 405-410.

Lafferty, KD & Kuris, AM 1994, in NEMO proceedings of the conference and workshop Seattle, 1993.

Lafferty, KD & Kuris, AM 1996. Biological control of marine pests. Ecology 77 (7), pp. 1989-2000.

Laiho, KU 1998. Infected ballast water: Heat treatment may not work with modern cooling systems, but will ultrasonics? MER-Marine Engineers Review, pp. 45

Laughton, R, Moran, T & Brown, G 1992. A Review and Evaluation of Ballast Water Management And Treatment Options to Reduce the Potential for the Introduction of Non-native Species to the Great Lakes. 45 pp.

http://www.pollutech.com/papers/p22.htm

Locke, A, Reid, DM, Sprules, WG, Carlton, JT & van Leeuwen, HC 1991. Effectiveness of mid-ocean exchange in controlling freshwater and coastal zooplankton in ballast water. Canadian Technical Report of Fisheries and Aquatic Sciences No. 1822. 98 pp.

Locke, A, Reid, DM, van Leeuven, HC, Sprules, WG & Carlton, JT 1993. Ballast water exchange as a means of controlling dispersal of freshwater organisms by ships. Can. J. Fish. Aquat. Sci. 50, pp. 2086-2093.

McCarthy, SA & Khambaty, FM 1994. International dissemination of epidemic Vibrio cholerae by cargo ship ballast and other nonpotable waters. Applied and Environmental Microbiology 60 (7), pp. 2597-2601

Medcof, JC 1975. Living marine animals in ships' ballast water. Proc. Natl. Shellfish Ass. 65, pp. 54-55 or 11-22.

Middleton, MJ 1982. The Oriental Goby, Acanthogobius flavimanus (Temminck and Schlegel), an introduced fish in the coastal waters of New South Wales, Australia. J. Fish. Biol. 21, pp. 513-523.

Morgan, G, Wright, D, Dawson, R, Bissing, A, Crawford, D, Barrcato, J & Coogan, J 1999. UV disinfection of ballast waters: Effects of organism size on system scaling. First National Conference on Marine Bioinvasions, January 1999, Cambridge, Massachusetts (summary only, p. 102).

Oemcke, DJ & van Leeuwen, H 1998. Potential of ozone for ballast water treatment. CRC Reef Research Centre and the Ports Corporation of Queensland, 40 pp.

Oemcke, D. 1999. The treatment of ships’ ballast water. EcoPorts Monograph Series No. 18 (Ports Corporation of Queensland, Brisbane), 102 pp.

Parsons, MG & Harkins, RW 1999. Filtration mechanical test results from the Great Lakes Ballast Demonstration Project. First National Conference on Marine Bioinvasions, January 1999, Cambridge, Massachusetts (summary only, p. 114).

Pierce, RW, Carlton, JT, Carlton, DA & Geller, JB 1997. Ballast water as a vector for tintinnid transport. Marine Ecology Progress Series 149, pp. 295-297.

Pini, RC 1999. Ballast water purified with UV waves. Photonics Spectra 33 (3), pp. 30-31.

Prior, AD 1995. Ballast water exchange study: Phase I, A study for the Transport Canada Marine Regulatory Directorate by Melville Shipping, Ottawa Ontario Canada.

Reid, DM & Carlton, JT 1997. Shipping study I-A. A study of the introduction of aquatic nuisance species by vessels entering the Great Lakes and Canadian waters adjacent to the United States. Report No. CG-D-17-97, U.S. Coast Guard, Groton CT and U.S. Dept. of Transportation, Washington, DC.

Rigby, G, Steverson, I & Hallegraeff, G 1991. Environmental problems and treatment options associated with the international exchange of shipping ballast waters. The Nineteenth Australian Chemical Engineering Conference, Newcastle, Australia, CHEMECA 91, pp. 221-230.

Rigby, G & Hallegraeff, G 1992/1993. Ballast water exchange trials and marine plankton distribution on the M.V. "Iron Whyalla". AQIS Ballast Water Research Series 2, 35 pp.

Rigby, G & Hallegraeff, G 1992/1993. Ballast water exchange trial and behaviour of marine plankton on the M.V. "Iron Whyalla", in Japan and en-route to Canada. AQIS Ballast Water Research Series 2, 44 pp.

Rigby, G, Steverson, I, Bolch, CJ & Hallegraeff, G 1993. The transfer and treatment of shipping ballast waters to reduce the dispersal of toxic marine dinoflagellates. In: Smayda, TJ & Shimizu, Y (eds). Toxic Phytoplankton Blooms in the Sea. Elsevier Science Publishers B.V.

Rigby, G & Hallegraeff, G 1994. The transfer and control of harmful marine organisms in shipping ballast water: behaviour of marine plankton and ballast water exchange trials on the MV "Iron Whyalla". Journal of Marine Environmental Engineering 1, pp. 91-110.

Rigby, GR, Taylor, AH, Hallegraeff, GM & Mills, P. 1995. Progress in research and management of ships’ ballast water to minimise the transfer of toxic dinoflagellates. In: Lassus et al (eds). Harmful Marine Algal Blooms. Lavoisier, Intercept Ltd.

Rigby, G, Hallegraeff, G & Sutton, C 1997. Ballast water heating and sampling trials on the BHP ship M.V. Iron Whyalla in Port Kembla and en-route to Port Hedland. AQIS Ballast Water Research Series 11, 40 pp.

Rigby, G, Hallegraeff, G & Sutton, C 1998. Ballast water heating and sampling trials on the BHP ship M.V. Iron Whyalla en-route from Mizushima to Port Hedland. AQIS Ballast Water Research Series 11, 19 pp.

Ryle, M 1995. Ballast exchange needs utmost care. The Motor Ship, pp. 30-32.

Simberloff, D & Stiling, P 1996. Risks of species introduced for biological control. Biological Conservation 78, pp. 185-192.

Spurgeon, J. 1999: The socio-economic costs and benefits of coastal rehabilitation and creation. Mar. Poll. Bull. Vol. 37. no. 8-12.

Subba Rao, DV, Sprules, WG, Locke, A & Carlton, JT 1994. Exotic phytoplankton from ships’ ballast waters: Risk of potential spread to mariculture sites on Canada’s east coast. Canadian Data Report of Fisheries and Aquatic Sciences 937, 51 pp.

Sutton, CA, Murphy, K, Martin, RB & Hewitt, CL 1998. A review and evaluation of ballast water sampling protocols. CRIMP at CSIRO 18, 113 pp.

Wathern, P. (1988). Environmental Impact Assessment. Theory and Practice. Routledge, London.

Williams, RJ, Griffiths, FB, van der Wal, EJ & Kelly, J 1988. Cargo vessel ballast water as a vector for the transport of non-indigenous marine species. Estuarine, Coastal and Shelf Science 26, pp. 409-420.

Wonham, MJ, Walton, WC, Frese, AM & Ruiz, GM 1996. Transoceanic transport of ballast water: biological and physical dynamics of ballasted communities and the effectiveness of mid-ocean exchange. Report for U.S: Fish and Wildlife Service, Washington, DC.

Woodward, JB, Parsons, MG & Troesch, AW 1994. Ship operational and safety aspects of ballast water exchange at sea. Marine Technology 31 (4), pp. 315-326.

Zhang, F & Dickman, M 1999. Mid-ocean exchange of container vessel ballast water. 1:Seasonal factors affecting the transport of harmful diatoms and dinoflagellates. Marine Ecology Progress Series 176, pp. 243-251.