Content uploaded by Guillaume Drillet
Author content
All content in this area was uploaded by Guillaume Drillet on Nov 08, 2016
Content may be subject to copyright.
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
E
MARINE ENVIRONMENT PROTECTION
COMMITTEE
70th session
Agenda item 4
MEPC 70/INF.21
17 August 2016
ENGLISH ONLY
HARMFUL AQUATIC ORGANISMS IN BALLAST WATER
Same risk area approach to exemptions under regulation A-4
of the Ballast Water Management Convention
Submitted by Singapore
SUMMARY
Executive summary:
This document presents information supporting the same risk area
concept, as referred to in document MEPC 70/4/8
Strategic direction:
7
High-level action:
7.2.2
Output:
7.2.2.4
Action to be taken:
Paragraph 6
Related documents:
MEPC 47/2; MEPC 48/2; MEPC 49/INF.6; MEPC 67/2/12,
MEPC 67/INF.23; MEPC 68/12/2; MEPC 69/4/11, MEPC 69/INF.25;
MEPC 70/4/8; PPR 2/5/3 and PPR 2/21
Introduction
1 As noted in document MEPC 70/4/8, regulation A-4 of the BWM Convention contains
provisions for multiple "Parties" to grant exemptions for multiple "ships" on multiple "voyages"
between multiple "specified ports or locations". However, the procedures for what is effectively
a regional approach to exemptions and the concepts underpinning such an approach, such as
the concept of a same risk area (SRA), have not been made clear.
2 At MEPC 68, the Committee agreed to re-establish the Ballast Water Review Group
at MEPC 69 to develop guidance on the "same risk area" focusing on short-sea shipping
(i.e. not intercontinental/ transoceanic voyages). Owing to time constraints at MEPC 69,
the Committee agreed to defer consideration of the matters of exceptions and exemptions
under the BWM Convention to MEPC 70.
MEPC 70/INF.21
Page 2
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Proposal
3 The annex
1
to this document contains information reported in a study entitled "A Study
on Same Risk Area – With Regards to Ballast Water Management Convention regulation A-4
on Exemptions to Ships" that aims to help discussions on the development of the SRA
guidance.
4 This information on SRA proposes a complementary approach to exemptions to
existing measures. It does not require any alteration of regulation A-4 or the current
Guidelines (G7). Furthermore, it clarifies the procedures under which an exemption may be
granted under regulation A-4 by multiple States for multiple voyages between multiple ports
and locations.
5 However, it is important to keep in mind that the study presented here does not
represent a full SRA assessment but merely offers a list of examples supporting the
methodologies that may be applied in a final study defining a SRA for the region. As such, the
final boundaries of an eventual SRA in the region should be verified through a detailed study.
Action requested of the Committee
6 The Committee is invited to note the information provided.
***
1
The boundaries and names shown and the designations used in the maps contained in the annex do not
imply official endorsement or acceptance by the International Maritime Organization.
MEPC 70/INF.21
Annex, page 1
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
ANNEX
A STUDY ON THE SAME RISK AREA CONCEPT
A Study on Same Risk Area
With Regards to Ballast Water Management
Convention Regulation A-4 on Exemptions to Ships
Final Report
Project Manager
Justine Saunders
Project Director
Dr Guillaume Drillet
Quality Supervisor
George Foulsham
This report has been prepared under the DHI Business Management System
certified by Bureau Veritas to comply with ISO 9001 (Quality Management)
MEPC 70/INF.21
Annex, page 2
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Executive Summary
The International Convention for the Control and Management of Ships' Ballast Water
and Sediments (The Ballast Water Management Convention or BWMC) (IMO, 2004)
and its associated guidelines aim to reduce the impact of potentially harmful aquatic
organisms and pathogens by preventing their spread from one region to another, by
establishing standards and procedures for the management and control of ships'
ballast water and sediments.
Once the BWMC is in force, most ships will be required to treat their ballast water
using on-board type approved Ballast Water Management Systems (BWMSs)
applying physical and/or chemical means to meet the ballast water discharge criteria
(regulation D-2 of the BWMC). In effect, this means that the potential primary invasion
from inter-oceanic voyages will be controlled through BWMS. However, ship(s) on
short-sea voyage(s) between specified ports or locations across international borders
may be granted an exemption from applying ballast water management systems
under BWMC regulation A-4, if it is decided that the risk of transfer of invasive species
is acceptable. This implies that a risk assessment should be carried out and Guideline
(G7) details the recommended process for this. Furthermore, regulation A-4 allows an
exemption to be granted for multiple ships and voyages between specified ports and
locations. This supports a regional approach to exemption.
Although low risk and high risk scenarios are described in evaluating the risk from the
transfer of invasive species via ballast water, the risk assessment approaches
recommended do not take into account the ability for species to disperse through
natural mechanisms. So the risk from shipping is not set into the context of the natural
baseline. In effect, the principle of proportionality in risk management has not been
considered in G7.
Mobile aquatic species and pelagic life stages of marine organisms may disperse
naturally across international borders, irrespective of other vectors of transfer such as
ballast water. It follows that ships that take short-sea voyages within such an area of
natural dispersion are unlikely to greatly alter the consequences from the transfer of
potentially harmful and invasive species. Based on rates and patterns of natural
dispersion, the area can be viewed as a "Same Risk Area" or SRA.
The study presented here proposes a complementary approach for exemption to that
recommended in G7, for multiple State, short-sea shipping based on the concept of
SRA. It is offers a regional and proportionate approach to exemption that supports
consistency, transparency and efficiency in the regulatory process while still providing
the same level of environmental protection relevant to the degree of risk.
The present report offers:
1. A chronological review of the SRA approach in the light of discussions taking
place at the International Maritime Organisation (IMO) and elsewhere;
2. An explanation of the rationale behind the use of a SRA approach for the
implementation of the A-4 regulation (Exemption) of the Ballast Water
Management Convention
3. A series of regional examples supporting the proposed methodology to carry
out risk assessment using the SRA approach in Southeast Asia;
MEPC 70/INF.21
Annex, page 3
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
4. A theoretical comparison of different risk assessment methods presenting
their advantages and disadvantages/weaknesses; and
5. A proposed workflow based on scientific principles for the implementation of
the SRA approach may it be recognized and accepted by other States at
IMO.
MEPC 70/INF.21
Annex, page 4
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
CONTENTS
1 Background of the Study ................................................................ 5
2 Introduction to the Same Risk Area Approach in the Southeast
Asian context ................................................................................... 8
2.1 The SRA approach with respect to Regulation A-4 and G7 ..................................... 8
2.2 Mitigation Measures ................................................................................................ 11
2.3 Existing areas of 'similar risk' or 'low risk' within and in between national waters .. 12
2.3.1 Designated Ballast Water Exchange Areas ............................................................ 12
2.3.2 United States Coast Guard ..................................................................................... 12
2.3.3 United States state laws ......................................................................................... 13
2.4 Conclusion .............................................................................................................. 13
3 Concept underpinning the SRA Approach .................................. 14
3.1 The concept of risk .................................................................................................. 14
3.2 The concept of short-sea shipping .......................................................................... 15
3.3 The concept of biogeographic regions .................................................................... 15
3.4 The concept of marine connectivity ........................................................................ 17
3.5 The dispersal of species ......................................................................................... 18
3.5.1 Examples of dispersal evaluations.......................................................................... 19
3.5.2 Approach to evaluate the dispersal of organisms ................................................... 19
4 Evidence for Connectivity in the inner EAS ................................ 20
4.1 Introduction to the EAS ........................................................................................... 21
4.2 Regional connectivity .............................................................................................. 22
4.2.1 UNEP/GEF South China Sea Project ..................................................................... 22
4.2.2 Flushing rates and residence times of water bodies in the Sunda Shelf ................ 24
4.2.3 Particle-tracking analysis ........................................................................................ 28
4.3 Local -scale connectivity ......................................................................................... 38
4.4 Conclusions for the EAS Region............................................................................. 38
4.4.1 Coral larvae dispersion ........................................................................................... 39
4.4.2 Sea-Star dispersion ................................................................................................ 45
5 General Methodological Approach to Assessing a SRA ............ 51
5.1 Draft Procedures ..................................................................................................... 51
5.2 Define Target Species for SRA ............................................................................... 52
5.2.1 Methods .................................................................................................................. 52
5.2.2 Findings ................................................................................................................... 54
5.2.3 Conclusions ............................................................................................................. 55
5.3 Hydrodynamic modelling and Agent-Based Modelling ........................................... 55
5.3.1 HD model ................................................................................................................ 56
5.3.2 ABM ........................................................................................................................ 56
6 Comparison of Protection Levels ................................................. 58
6.1 Methodology for comparison ................................................................................... 59
6.2 Scenarios ................................................................................................................ 60
6.3 Comparison of Risk Assessment Principles ........................................................... 62
6.4 Conclusion .............................................................................................................. 67
7 Proposed Guidelines for Same Risk Area ................................... 68
8 References ..................................................................................... 74
MEPC 70/INF.21
Annex, page 5
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
1 Background of the Study
The introduction of exogenous species to new ecosystems can evolve towards
uncontrollable situations where the introduced species become invasive; this occurs in
approximately 10% of the introductions (Boudouresque and Verlaque, 2002). Invasive
species are viewed as a major threat to aquatic ecosystems and have been reported to
affect the global economy as well. In the USA alone, the impact of aquatic invasive
species is estimated to range between millions and billions of dollars annually (Lovell
et al., 2006). The shipping industry has been identified as a major source of transport
of species across ecosystems with about a third of the introductions due to fouling on
the ship hulls and another third due to ballast water exchanges (Gollasch, 2007). Ballast
water is therefore an issue of concern which has raised a tremendous amount of
attention in the last decades (Eno et al., 1997, Ruiz et al., 1997, Carlton, 1993, Seebens
et al., 2013).
To address this issue, the International Convention for the Control and Management
of Ships' Ballast Water and Sediments (Ballast Water Management Convention or
BWMC) was adopted by IMO in 2004 (IMO, 2004). The BWMC and its associated
guidelines aim to reduce the impact of potentially harmful aquatic organisms and
pathogens by preventing their spread from one region to another, by establishing
standards and procedures for the management and control of ships' ballast water and
sediments. The BWMC is to enter into force exactly one year after at least 30
countries representing 35% of the world merchant shipping tonnage have ratified
(Article 18). Based on the recent ratifications and announcement of interest from
additional countries, it is expected that the convention will be fully ratified in 2016 and
will therefore enter into force in 2017.
Article 3.2 of the Convention specifies six cases for which the Convention shall not apply:
a) Ships not designed or constructed to carry ballast water;
b) Ships of a Party which only operate in waters under the jurisdiction of that Party,
unless the Party determines that the discharge of ballast water from such ships
would impair or damage their environment, human health, property or
resources, or those of adjacent or other States;
c) Ships of a Party which only operate in waters under jurisdiction of another Party,
subject to the authorisation of the latter Party for such exclusion. No Party shall
grant such authorisation if doing so would impair or damage their environment,
human health, property or resources, or those of adjacent or other States. Any
Party not granting such authorisation shall notify the Administration of the ship
concerned that this Convention applies to such ship;
d) Ships which only operate in waters under the jurisdiction of one Party and on
the high seas, except for ships not granted an authorisation pursuant to
subparagraph (c), unless such Party determines that the discharge of ballast
water from such ships would impair or damage their environment, human
health, property or resources, or those of adjacent or other States;
e) Any warship, naval auxiliary or other ship owned or operated by a State and
used, for the time being, only on Government non-commercial service.
However, each Party shall ensure, by the adoption of appropriate measures not
impairing operations or operational capabilities of such ships owned or operated
by it, that such ships act in a manner consistent, so far as is reasonable and
practicable, with this Convention; and
MEPC 70/INF.21
Annex, page 6
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
f) Permanent ballast water in sealed tanks on ships that is not subject to
discharge.
Eventually, most ships will be required to treat their ballast water using on-board type
approved Ballast Water Management Systems (BWMSs) using physical and/or
chemical means to meet the ballast water discharge criteria (regulation D-2). In effect,
this means that the potential primary invasion from inter-oceanic voyages will be
controlled through BWMS.
The BWMC by definition only regulates ships exchanging ballast water across
international borders, and therefore domestic voyages are excluded from the
provisions of the Convention. Ships travelling between ports in the same country are
under no obligation to comply with the BWMC regardless of the distance covered.
This convention therefore creates a scenario where exchange of ballast water
between distant ports of a single country may be unregulated (if not regulated at the
national level), while the discharge of ballast water between ports in neighbouring
countries (for example across a strait) is subject to the regulations set out by the
BWMC despite the expected higher ecosystem similarity at the local scale. This is
biologically unsound and was highlighted in a report to the Danish Nature Agency
(Stuer-Lauridsen and Overgaard, 2014):
"You are obliged to meet the D-2 Standard following the 4 km journey
between Elsingør in Denmark and Helsingborg in Sweden, where intense
shipping traffic has taken place for 100+ years, but not if you travel 1,400 km
from Swedish Luleå to Helsingborg or for that matter the 600 km from Danish
Esbjerg in the North Sea to Elsinore in The Sound. Even without a G7 risk
assessment at hand it is likely that the risk associated with the 4 km crossing
of the Sound may indeed pose the lowest risk; yet, this is the only journey
requiring a risk assessment."
The BWMC is meant to be an international instrument and therefore its regulations
are not to interfere with any additional regulations that may be implemented at national
levels. However, Article 13.3 supports the endeavour of regional collaboration:
"In order to further the objectives of this Convention, parties with common
interest to protect the environment, human health, property and resources in
a given geographical area, in particular those parties bordering enclosed or
semi enclosed seas, shall endeavour, taking into account characteristic
regional features, to enhance regional cooperation, including through the
conclusion of regional agreement consistent with this Convention. Parties
shall seek to co-operate with the parties to regional agreements to develop
harmonized procedures."
The BWMC provides some flexibility in terms of the application of its regulations.
According to regulation A-3.5 (Exceptions) in the Annex to the BWMC, the
requirements of regulation B-3 which is replaced by the Assembly
Resolution A 1088(28) shall not apply to:
"the discharge of ballast water and sediments from a ship at the same
location where the whole of that ballast water and those sediments originated
and provided that no mixing with unmanaged ballast water and sediments
from other areas has occurred."
Regulation A-4 (Exemptions) offers the opportunity "to a ship or ships on a voyage or
voyages between specified ports or locations" presenting an acceptable low risk of
MEPC 70/INF.21
Annex, page 7
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
the transfer and introduction of harmful aquatic organisms and pathogens to be
granted an authorisation not to comply with regulation B-3 of the BWMC; an
exemption. This exemption may be granted by authorities after the evaluation of risks
following G7 Guidelines developed by the IMO.
In effect, an exemption if granted allows the flexibility for ships to exchange ballast
water between specified international ports. This partly answers the challenges that
may occur in the case of short distance international voyages, for example between
Singapore and Indonesia in the Southeast Asian context and as given in the above-
mentioned example of Sweden/Denmark. Considering that the cost of a risk
assessment is estimated to be cheaper than fitting or retrofitting a type-approved
BWMS on-board a ship, the flexibility offered by regulation A-4 and the associated G7
Guidelines has been considered sufficient. However, a limitation of the exemption
procedure is that it restricts the exemption for ships only between the specific ports
for which the risk assessment is carried out, hampering any possibility of route
changes. In ports and countries where limited biological and physical data exists, the
emerging protocols available for conducting biological sampling and monitoring as a
basis for risk assessment may become so prescriptive that they effectively remove
the possibilities offered by regulation A-4. The cost of generating risk assessments
for every port-to-port route of every relevant ship would be prohibitive to the shipping
industry as a whole.
Two of the principles for risk assessment in G7 are transparency and consistency.
However, the terminology used in the papers submitted to MEPC/PPR, in the
Convention, its annexes and its guidelines has been generating confusion among the
countries represented at IMO. There has been for instance references to the following:
geographical area
same location
specific ports or locations
other areas
same hydrographical regime (footnote 19 of the annex to document
MEPC 48/2
biogeographic region which is defined in the G7 Guidelines as "A large
natural region defined by physiographic and biologic characteristics within
which the animal and plant species show a high degree of similarity. There
are no sharp and absolute boundaries but rather more or less clearly
expressed transition zones
biogeographical scheme (also referred to in G7 paragraph 6.2.3).
In January 2014, the Danish Partnership for Ballast Water agreed to commission a
study on "Same Risk Area" in connection with the implementation of the BWMC and
two subsequent papers were submitted to MEPC 67/INF.23 by Denmark and
Interferry. Following the submissions MEPC 67/2/12, MEPC 67/INF.23 by Denmark
and Interferry, and the submission PPR 2/5/3 by Croatia, Denmark, Singapore, ICS
and Interferry, the conclusions of the PPR 2 meeting (reported in PPR 2/21)
established that a new guidance document or a revision of Guidelines (G7) could be
an appropriate way forward, but concrete proposals were needed.
MEPC 70/INF.21
Annex, page 8
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Denmark and Interferry submitted an additional paper on SRA for consideration at
MEPC 69 (MEPC69/4/11 as well as a full technical paper MEPC 69/INF.25).
A definition of SRA was provided by MEPC 69/INF.25 as "A body of water
characterized by the an (sic) equal risk level from natural dispersal of target species".
This working definition of a SRA is helpful in that it encompasses the different above-
mentioned terminologies presented in the different documents used by member
states at IMO. Although not explicit in the definition, the supporting document (MEPC
69/INF.25) clarifies that the risk level from natural dispersal of target species should
be compared against the risk regarding transfer via ballast water; the added risk from
exchange of ballast water must be acceptable.
An initial response to the proposal was provided by Canada (MEPC 69/4/15),
essentially supporting the scientific rationale behind the concept, but with suggestions
on additional considerations (these are addressed later in this current paper). The Ballast
Water Review Group was established at MEPC 69 and was instructed to develop
guidance on the "same risk area", on the basis of the annex of MEPC 69/4/11. However,
due to time constraints, consideration of the matters were deferred to MEPC 70.
In summary, the concept of "Same Risk Area" is broadly accepted in principle and
detailed guidance on procedures for Exemption under this approach is expected to
be developed in the near future and may be incorporated as an additional guidance
to the existing guidelines either as a new risk assessment approach or as a variation
of an existing approach.
2 Introduction to the Same Risk Area Approach in the Southeast Asian context
2.1 The SRA approach with respect to Regulation A-4 and G7
We define SRA in this paper as "An area delimited by the high probability of natural
spread of target species that potentially present a risk of bio-invasion via ballast
water". This is in contrast to the working definition in MEPC 69/INF.25 in that there is
no comparison of risk levels from different vectors, whether natural or anthropogenic.
The SRA approach defines a specific area (based on specified ports and locations)
within which all ships travelling solely within the boundaries of this area can be
exempted from ballast water management obligations because ballast water
exchanges result in the same consequences from target species transfer as the
estimated natural dispersal over time.
The SRA approach is intended to be applied in a situation where a limited area is
served by short-sea shipping (see definition in Section 4.2) and includes several ports
in two or more countries in close proximity to each other. In line with other exemptions
under regulation A-4, the exemption granted on the basis of the SRA concept should
only be granted to vessels operating exclusively in the defined area and not mixing
their ballast waters and sediments with water and sediments originating from outside
that area. The SRA approach recognizes that the potential primary invasion from inter-
oceanic voyages (e.g. from the eastern tropical Atlantic to Southeast Asia) will be
controlled through BWMS.
Regulation A-4 allows for a regional approach to exemption in providing for an
assessment of multiple "ships" and "voyages" between "specified ports or locations"
that constitute the defined area of similar risk. For countries with regional/local traffic
MEPC 70/INF.21
Annex, page 9
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
calling at several ports across borders, this would result in reduced administrative
burdens. For the shipping industry, this approach could reduce costs and benefit the
movement of cargo in short-sea traffic.
Regulation A-4 also stipulates that an exemption may only be granted "based on the
guidelines on risk assessment" in G7. Under G7, three risk assessment methods are
currently recommended in relation to granting an exemption in accordance with
regulation A-4:
Environmental matching risk assessment
Species biogeographical risk assessment
Species-specific risk assessment
The environmental matching risk assessment relies on comparing environmental
conditions between locations that are distant but share similar environmental
parameters (for example, salinity ranges and temperature). If two distant locations
shared the same environmental conditions (e.g. a location in the eastern tropical
Atlantic and a location in Southeast Asia) then the risk of transfer and successful
settlement of organisms from one to the other is high. This is an acceptable approach
for the development of a risk assessment of the discharge or organisms that are from
very distant origins and therefore this approach cannot be used within the context of
a SRA approach. However, Stuer-Lauridsen and Overgaard (2014) suggests that this
is a useful approach to use in defining a Same Risk Area as similar environmental
conditions (i.e. salinity and temperature) in two closely located ports are an indication
of the ports sharing the same risk profile.
The species' biogeographical risk assessment approach is based on the comparison
of the overlap of native and non-indigenous species to evaluate environmental
similarity and to identify high risk invaders. Under the premise of the SRA concept, an
overlap in species is to be expected given that the ports or locations are naturally
connected. Indeed, it would be of more concern if the sites did not share the same
species.
Species-specific risk assessment also evaluates the distribution and characteristics
of individual target species, comparing their characteristics with the environmental
conditions within the ports in question to determine the likelihood of transfer and
survival. In the context of an SRA approach to exemption and considering that it is
intended for short-sea shipping traffic, the States undergoing such a study would need
to identify and agree on the target species to be studied before the risk assessment
is undertaken thereby applying the guidelines under the species-specific risk
assessment. This effectively equals the first two steps in a species-specific risk
assessment: identification of target species and simple risk assessment.
The existing G7 Guidelines consider ports as a source and a receptor of potentially
harmful and invasive species (or target species) but in the evaluation stage eventually
ignore the potential dispersal characteristics of such species and the natural
connectivity between ports waters and adjacent coastal waters. A practical approach
to consider when implementing the BWMC should include an understanding of the
'baseline' risk that potentially invasive species may naturally extend across state
borders, based on their biology, their dispersal patterns and their habitat
requirements. A "Same Risk Area" is described above as a body of water
characterized by an equal risk level from the natural dispersal of target species. In
this sense, ports in an SRA are no longer labelled as donor and recipient ports as
defined under G7, as they perform both functions under natural two-way processes
MEPC 70/INF.21
Annex, page 10
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
of dispersion. Definitions of recipient and donor ports can become less relevant on a
larger spatial and temporal scale when we factor in natural connectivity.
It is not the size of the port that is important as a donor of potentially invasive species
but the degree to which it is naturally connected to neighbouring ports and locations.
This is particularly relevant for the Port of Singapore which is one of the largest ports
in Southeast Asia. However, lying at the junction between the South China Seas,
Malacca Straits and Java Sea, Singapore is also a centre of natural aggregation of
species (i.e. recipient port). UNEP, (2004a) noted that the South China Seas has
received species from both the Pacific and Indian Ocean basins.
The boundaries of the SRA is based on ecological processes and therefore may
extend across two or more state borders. This is the natural or baseline boundary of
risk for target species.
The method used for conducting a risk assessment must be accepted by the States
it concerns and eventually any neighbouring States which may be affected by such
agreements. In paragraph 8.1 of the G7 Guidelines, it is stipulated that:
In accordance with regulation A-4.3, Parties shall consult any State that may
be adversely affected from any exemptions that may be granted. This should
include adjacent States and any other States that may be affected, including
those located in the same biogeographic region as the recipient port(s).
States should exchange information and endeavour to resolve any identified
concerns. Sufficient time must be given for affected States to consider
proposed exemptions carefully.
This is supported by the SRA approach proposed here in that agreement is sought
among States involved in the SRA on aspects such as the target species identified. It
is recognized that neighbouring States outside of the SRA may still need to be
consulted with if they are likely to be affected by the proposal.
The SRA approach represents a proportionate approach to exemption that recognizes
the potential low risk presented by short-sea shipping voyaging through a body of
water (shared by multiple States) clearly defined by hydrodynamic properties and
where species of interest are naturally dispersed over time. In general, measures
applied to provide a high level of protection of the environment, taking in account the
precautionary principle, should neither be disproportionate to the degree of risk, nor
discriminatory (Tridimas, 2013, Veinla, 2004)
2
. With respect to a SRA, the costs of
installing a BWMS may be disproportionate to the benefits, i.e. in a SRA the benefits
may be minimal if the target species can naturally disperse over the SRA extent.
Although regulation A-4 allows for flexibility in the approach for exemption, G7 does
not recognize the concept of proportionality under the risk assessment principles. The
risk assessments do however allow for some flexibility. For example, under the
species-specific risk assessment a simple assessment may be considered
appropriate or a more detailed assessment that includes the likelihood of survivorship
of target species in ballast water of the specific voyage may be necessary. Under the
evaluation stage of the risk assessment, different levels of risk scenarios may be
recognized. However, for the species-specific risk assessment only a high risk
scenario is described. SRA could be recognized as an appropriate example of a low
risk scenario under the species-specific approach; the SRA effectively delimits an
area of potential low risk from the discharge of ballast water (see para. 6.5.5 of G7).
2
See also Case Law example: Case 302/86 Commission v. Denmark, European Court Review 1988, 4607.
MEPC 70/INF.21
Annex, page 11
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Under the evaluation stage of a risk assessment, it is recommended that a risk
assessment should consider the overall probability of a successful invasion of target
species via ballast water based on the number of organisms transported via ballast
water and the frequency with which they are introduced. It is suggested that this is
estimated from information on the total volume of ballast water exchanged; the volume
of water discharged in any event (voyage); the total number of discharge events; and
the temporal distribution of discharge events. Under the SRA approach, detailed
information on the volumes of ballast water exchanges in the defined area are not
required. This makes the SRA approach applicable to all ships in the area, whatever
their ballast water exchange activities.
Finally, there are the timeframes for the validity of the assessment to consider.
Exemptions made under the SRA will be valid for five years, as is specified under
regulation A-4 and the current G7 guidelines. The validity of the SRA itself (i.e. its
boundaries and specified locations and/or ports) should be considered based on the
review period selected and agreed by SRA States for the SRA (as per section 7;
para 10) or unless evidence arises to suggest that there have been changes in the
presence of target species and hydrographic baselines of the SRA.
2.2 Mitigation Measures
The convention has dispositions about measures to take in the event of a bloom of
potentially harmful organisms occurring in waters where ship may ballast water from
(regulation C-2). This should apply for both ships using BWMS and ships under an
exemption.
"A Party shall endeavour to notify mariners of areas under their jurisdiction
where ships should not uptake Ballast Water due to known conditions. The
Party shall include in such notices the precise coordinates of the area or
areas, and, where possible, the location of any alternative area or areas for
the uptake of Ballast Water"
However, additional mitigation measures should be considered for any exemption
under regulation A-4 if a non-native, unassessed species turns up at any time,
regardless of whether a Same Risk Area or individual port-to-port approach to
exemption is adopted. A new species may be identified through a physical survey
conducted for any reason (commercial, port-related or academic research). A
suitable approach in such a situation might involve an evaluation of the risk from the
new species at the earliest. This should include an assessment of the natural
dispersal capabilities of the species, taking into account population dynamics and
environmental requirements.
The notification of the area affected by a new introduction will come under regulation
C-2. Mitigation measures should ensure that if even only one port is affected within
the SRA by the arrival of a new species that the exemption for the entire area is not
withdrawn. Unaffected states within the SRA should be able to carry on their activities
with other unaffected states as before until the results of the updated risk assessment
are available. However, in the event where the re-evaluation of risk carried out for the
newly introduced species reveals a non-acceptable risk, the SRA should be re-
evaluated and eventually exemption may be withdrawn. In such cases, the involved
state shall provide guidance on the timeline for ship-owners to prepare for compliance
with the D-2 regulation.
MEPC 70/INF.21
Annex, page 12
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
2.3 Existing areas of 'similar risk' or 'low risk' within and in between national waters
Existing agreements between Port States that are aligned with the concept of the SRA
(but maybe not its methodology) are in use already. Within single Port States as well,
existing national regulation supporting the concept of SRA are applied. A number of
examples were presented in MEPC 69/INF.25. In the following paragraphs, we
present some examples in more detail.
2.3.1 Designated Ballast Water Exchange Areas
As part of regulation B-4.2 of the BWMC, in sea areas where the distance from the
nearest land or the depth does not meet the parameters described in regulation B-4.1,
the port State may designate areas where a ship may conduct ballast water exchange
(BWE). G14 provides guidelines for designating such areas. The area is selected
based on the understanding of hydrodynamic patterns and bathymetry around the
ports, ensuring that the de-ballasted water is "flushed away" from sensitive areas where
potential invasions may occur. Areas to be avoided include high nutrient areas, areas
known to contain outbreaks, infestations, or populations of Harmful Aquatic Organisms
and Pathogens and areas of pollution from human activities such as sewage outfalls.
Sensitive areas such as protected areas, key fishery resources, aquaculture farms are
also to be avoided. A risk assessment of these site selection factors is produced in order
to ensure that the selected location and size of the BWE area presents the least risk to
the aquatic environment, human health, property or resources.
2.3.2 United States Coast Guard
The Department of Homeland Security, under the US Coast Guard (USCG), has laid
down rules for the management of ballast water for all non-recreational vessels, US
and foreign, that are equipped with ballast tanks and operate in the waters of the US,
in excess of 79 feet in length (Lloyd's Register, 2015). To be exempt from the Ballast
Water Management requirements, reporting and recordkeeping, a vessel is required
to meet one of the following criteria (Department of Home Security (US), 2012):
Crude oil tankers engaged in coastwise trade.
Vessels that do not operate outside the EEZ… must operate exclusively
within one Captain of the Port (COTP) zone in order to be exempt from
meeting the ballast water discharge standard.
To be exempt from the Ballast Water Management requirements, but not the reporting
and recordkeeping regulations, a vessel is required to meet one of the following criteria:
Seagoing vessels that operate in more than one COTP Zone, do not
operate outside of the Exclusive Economic Zone (EEZ), and are less
than or equal to 1,600 gross register tons or less than or equal to 3,000
gross tons (International Convention on Tonnage Measurement of
Ships, 1969).
Non-seagoing vessels.
Vessels that take on and discharge ballast water exclusively in one
COTP Zone.
MEPC 70/INF.21
Annex, page 13
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
2.3.3 United States state laws
Under Californian law, vessels are exempt from ballast water management if the
ballast water is discharged at the "same location" where the ballast water originated,
and proof can be shown that the water was not mixed with ballast water taken on in
an area other than mid-ocean waters. "Same location" is defined as within one
nautical mile (6,000 ft.) of the berth or within the recognized breakwater of a California
port or place at which the ballast water was loaded (California State Lands
Commission, 2015a).
In addition, two other regions within the Pacific coast region have been identified as
being of the "same location". In the first instance, all ports and places in the San
Francisco Bay area east of the Golden Gate Bridge, including the ports of Stockton
and Sacramento, are construed as the same California port or place. Also, the ports
of Los Angeles, Long Beach, and the El Segundo marine terminal are considered as
the same California port or place (California State Lands Commission, 2015b). A
result of such an exemption means that a vessel can transfer unmanaged ballast
water between, for example, Oakland and Sacramento, ports that are more than 50
nm apart (David et al., 2013).
In New York state, ballast water management regulations do not apply to vessels that
(United States EPA, 2009):
operate exclusively in the Great Lakes - St. Lawrence Seaway System
upstream of a line drawn from Cap-de-Rosiers to West Point, Anticosti
Island and then to the north shore of the St. Lawrence River along a
meridian of longitude 63 degrees West
operate exclusively within the waters of New York Harbour and Long
Island Sound; or
enter New York waters from ports of call within New Jersey and
Connecticut waters which are included in the definition of "waters of New
York Harbour and Long Island Sound" provided that the vessel has met the
requirements of this condition before entering the waters of New York
Harbour and Long Island Sound.
In Washington State, vessels are exempt from the requirements of ballast water
exchange if the ballast water and sediments originate solely within the waters of
Washington state, the Columbia river system, or the internal waters of British
Columbia south of latitude fifty degrees north, including the waters of the Straits of
Georgia and Juan de Fuca (Washington Department of Fish & Wildlife, 2009).
2.4 Conclusion
The current limitation of G7 is that it is a port-to-port risk assessment that does not
take into account natural dispersion of organisms, particularly between neighbouring
countries. There is a need therefore for an approach to exemption that addresses the
common interests of several states in an area and takes into account the levels of risk
from natural dispersion of target species. Some (large) states already apply solutions
to this issue through the identification of areas of 'similar risk' and 'low risk' or 'same
location'. These examples provide support for the integration of the SRA concept in
applying for exemptions. There is also provision for an area-based approach in the
current regulations in that they allow exemptions for multiple ships, voyages and ports
MEPC 70/INF.21
Annex, page 14
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
or locations. However, a scientific method for this integration needs to be detailed that
is rigorous and robust.
The Same Risk Area approach also answers some of the issues with terminologies
used in different documents at IMO in that it encompasses the concept of risk of
introduction of species which is underlining the Ballast Water Convention.
Finally, the present document proposes mitigation measures to deal with situations
where introduction of new species may occur in ports where exempted ship may
exchange ballast water.
3 Concepts underpinning the SRA Approach
3.1 The concept of risk
The GEF-UNDP-IMO Globallast Partnerships Programme and WMU (2013) defined
risk as "the probability that a hazard will lead to loss, or injury/damage to life, property
or the environment; it requires knowledge of the extent of exposure to the hazard
concerned".
Risk assessment is a scientific process that incorporates understanding on the
likelihood of an event occurring and its magnitude and consequences. Although risk
assessments can be qualitative or quantitative, risk assessments need to be
conducted in a systematic and rigorous manner if to be of value as a decision-making
tool.
The consequences of a risk should always be put into context of the natural system
or baseline. In an SRA with ports or locations that are naturally connected in all
directions through hydrodynamic and ecological processes, the consequence from
the transfer of potentially invasive species may have already occurred or is inevitable.
The introduction of a further vector of transfer such as ballast water exchange will not
change this consequence.
This brings to mind an English saying, "closing the stable door after the horse has
bolted", i.e. trying to stop something bad happening when it has already happened
and the situation cannot be changed. Similarly, with respect to the SRA concept, a
risk assessment of the introduction of potentially invasive species by ballast water
may not be necessary within the bounds of a SRA if it is illustrated in the baseline
assessment that species can and have transferred by natural mechanisms.
With respect to the concept of risk then, a "Same Risk Area" may therefore be
described as a body of water characterized by an equal risk level from the natural
dispersal of target species, irrespective of other vectors of transfer such as ballast water.
Risk assessment may also consider aspects of risk management. Risk management
refers to the suite of concrete responses to the risk, i.e. suppression, avoidance,
reduction and/or acceptance (OECD, 2003). Risk management should also involve
communication of risk to those affected. Importantly, the measures imposed to
manage the risk should be proportionate to the degree of risk posed by the activity. In
European Community law, the principle of proportionality has become a fundamental
principle which cannot be disregarded even in the choice of precautionary measures
(Veinla, 2004).
MEPC 70/INF.21
Annex, page 15
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
3.2 The concept of short-sea shipping
The term "short-sea shipping" is derived from European use where it is defined by the
European Commission (1999) as "the movement of cargo and passengers by sea
between ports situated in geographical Europe or between those ports and ports
situated in non-European countries having a coastline on the enclosed seas bordering
Europe." Short-sea shipping includes domestic and international maritime transport,
including feeder services along the coast, to and from the islands, rivers and lakes.
Feeder services form a short sea network between ports in order for the freight
(usually containers) to be consolidated or redistributed to or from a deep-sea service
in one of these ports. The concept of short sea shipping also extends to maritime
transport between the Member States of the Union and Norway and Iceland and other
States on the Baltic Sea, the Black Sea and the Mediterranean.
Alternatively, maritime traffic that crosses oceans is referred to as 'deep-sea shipping',
'intercontinental shipping' or 'ocean shipping'. While 'short-sea shipping' is the term
used in Europe, the United States also has used the term, 'marine highway', while
Australia, New Zealand and Canada have used the term "coasting trade" (Brooks, 2009).
In MEPC 69/INF.25 the application of a SRA with respect to short-sea shipping is
intended for "several ports in close proximity in two or more neighbouring countries -
e.g. along a coast line, across a strait, a bay or a river mouth."
In the context of Southeast Asia, we suggest that "short sea shipping" would similarly
relate to ships on short domestic and international routes along the coasts and among
the islands of the region. For example, this would include ships plying routes along
the southern coastline of SE Asia from northern Vietnam to Thailand, or traversing
the Malacca, Karimata and Singapore Straits, and crossing the Gulfs of Thailand and
Tonkin. This is of a comparable scale to the application of the concept in Europe.
3.3 The concept of biogeographic regions
Intuitively, to be able to apply the SRA concept, biogeographic regions resulting in an
acceptable low risk of the transfer of invasive species must be clearly defined.
MEPC 69/INF.25 suggests that an SRA would be expected to be defined in one
biogeographical region only. Various classifications of biologically defined areas
(e.g. ecoregions, bioregions) have been described previously (Olson et al., 2001,
World Wildlife Fund for Nature, 2008, Kelleher, 1996, etc ), although a majority of these
are heavily focused on the terrestrial environment or are limited in their spatial resolution.
Under G7, a biogeographic region is defined as "a large natural region defined by
physiologic and biologic characteristics within which the animal and plant species
show a high degree of similarity. There are no sharp and absolute boundaries but
rather more or less clearly expressed transition zones." Within the regulations of the
BWMC, the concept of large marine ecosystems (LMEs) developed by various
experts from the National Oceanic and Atmospheric Administration (NOAA) and the
University of Rhode Island (National Oceanic and Atmospheric Administration,
undated) has been embedded. Most LMEs are relatively large areas of ocean space
of approximately 200,000 km² or greater, adjacent to the continents in coastal waters
where primary productivity is generally higher than in open ocean areas. When
Sherman (1991) proposed the LME concept, he recognized that ecological and
physical connectivity, as opposed to political boundaries, are fundamental to the
management, functioning and conservation of the marine environment.
MEPC 70/INF.21
Annex, page 16
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
IMO-Globallast have also developed a "World Bioregion Map" which displays 204
discrete bioregions for use in displaying ballast water risk assessment results (Clarke
et al., 2003). The bioregions largely adhere to the system of marine protected areas
adopted by the International Conservation of Nature (IUCN) (Kelleher et al., 1995),
with changes made by the CSIRO-Centre for Research into Introduced Marine Pests
(Global Ballast Water Management Programme, 2003). The nearshore bioregions are
based on Kelleher et al. (1995) which were identified based on environmental and
biological characteristics by local marine experts within 18 separate ocean regions.
These were modified by Hewitt et al. (2002) for the purposes of assessing introduced
marine species distributions as follows:
Their offshore boundary was limited to 200 nautical miles, in order to
represent each country's coastal zone and continental shelf; and
Additional bioregions were added for oceanic islands and island
networks (e.g., Hawaiian Islands, Galapagos Islands, Canary Islands).
Hewitt et al. (2002) also assigned the nearshore bioregions to large-scale biological
provinces derived from Sherman (1991). These provincial associations represent
broad-scale patterns of transitions in biodiversity and species assemblages.
As an example, the East Asia Seas (EAS) contains 8 bioregions, of which the most
relevant to Southeast Asia are EAS-I to EAS-III and EAS-VI (Figure 3.1). EAS IV and
V are east of the Philippines and the island of New Guinea in the oceanic waters of
the Pacific. EAS VII and VIII are to the south of Indonesia in the Indian Ocean and
Timor Sea. Kelleher et al. (1995) also proposed second order subdivisions that are
more coastal in nature.
Figure 3.1 Boundaries of the IMO-Globallast bioregions in colour
Most of these schemes are based on biogeographical features of the coastal and shelf
waters, combining benthic and shelf pelagic (neritic) biotas. The boundaries are
based on taxonomic configurations, influenced by evolutionary history, patterns of
MEPC 70/INF.21
Annex, page 17
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
dispersal, and isolation. For example, range discontinuities, dominant habitats,
geomorphological features, oceanographic currents, and temperatures were all used
to identify areas and refine boundaries. However, there was little consideration of
nearshore hydrodynamic features that would be influencing port waters or coastal
areas of ballast water exchange. Definition of a Same Risk Area would follow similar
concepts although there would be more detailed consideration of the various scales
of hydrodynamic features, i.e. not just large-scale oceanographic ones.
3.4 The concept of marine connectivity
The understanding of "marine connectivity" as a concept is important in the SRA
approach proposed here. Several definitions of marine connectivity exist in the
literature (examples are provided below); however, it is most commonly used to
describe the exchange of individuals among marine populations. The larger the extent
at which (sub-) populations are connected, the larger the exchange of individuals and
the more connected the regions are (Cowen and Sponaugle, 2009). When
populations are totally unconnected, they may eventually evolve as different species
over time (speciation).
Molecular data and genetic markers are useful in determining genetic connectivity
between populations. Gene mutation, random genetic drift, and selection are responsible
for genetic diversification, while gene flow due to migration/transport of individuals is
responsible for the homogenisation of genetic data (Hedgecock, et al., 2007; Gagnaire
et al., 2015). In other words, a high exchange of individuals can allow for sufficient gene
transfer resulting in genetically similar populations. The opposite is true, as populations
tend to slowly diverge genetically if very low levels of exchange exist (Sale et al., 2010).
Data on genetic metrics, or more specifically the standardized allele-frequency variance,
can be used to make "qualitative predictions concerning the geographical extent to which
a non-native marine species will spread once established in a new area" (Palumbi, 2003).
Therefore, there is often a pretty robust correlation between the measured
connectivity between populations (molecular information) and the dispersal of
populations within a geographical area. The evaluation of connectivity between
populations informs scientists on the dispersal of species, whether through natural
dispersal or generated by the transport of individuals through human activities.
A key concept in the SRA approach is that connectivity must be multi-directional. This
was a point raised by Canada (MEPC 69/4/15): "an exemption decision based on the
SRA approach would appear to allow ballast transfers in any direction within the
SRA." If for example, modelling studies indicated a uni-directional flow of natural
dispersion from Port A to Port B but not in the reverse direction, then it would follow
that exemption could not be granted for a ship travelling between these ports. This of
course becomes more complex when more ports are added to the SRA. In effect, the
SRA study supporting the application for exemption must prove that all ports in the
SRA are connected to each other (see Figure 3.2 for an example).
MEPC 70/INF.21
Annex, page 18
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 3.2 Example of the multi-directional connectivity inherent in the SRA approach. All
three ports in the example must be connected, as illustrated by the arrows,
through natural dispersion.
3.5 The dispersal of species
The dispersal of species refers to the transport of individuals, populations and
eventually species across geographical areas, at all scales. It may be natural and
mediated by winds, currents, migrations, parasitism etc. or mediated by human
activities. The dispersal of species across ecosystems by human activities such as
the exchanges of ballast water is responsible for the introduction of exogenous
species in different ecosystems and is the source of invasion by non-indigenous
species. In the case of species introduction, the influence of human activities as a
vector is easy to determine though the activity responsible may sometimes be difficult
to clearly identify (Aquaculture, Shipping, Fishing, etc.). However, the extent human
activities are affecting the connectivity between populations of one area is more
unknown and more difficult to assess because is it concomitant with the natural
dispersal of organisms.
The natural dispersal of organisms is an important survival strategy as it ensures the
continued or increasing connectivity between populations and subsequently, securing
a larger gene pool. It also ensures the spread of offspring or shoots within the
ecosystems. The dispersal and the spread of a species is limited by the capacity of
the organism to use their environment as a vector. A benthic organism without a
planktonic phase has low capability to use the water column as a means of transport,
but mobile and planktonic organisms (or organisms with a planktonic life stage) would
be able to do so and therefore have the capacity to disperse further. The natural
dispersal of organisms in the aquatic environment is often linked to the life history
traits of organisms – those that live part of, or their entire life cycles, in the water
column (meroplankton and holoplankton, respectively
3
).
3
This assumes that most invasive species that are introduced through ballast water exchanges are planktonic
organisms because nekton is considered capable of avoiding the pumping of water in tanks.
MEPC 70/INF.21
Annex, page 19
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
3.5.1 Examples of dispersal evaluations
There are numerous examples in peer-reviewed literature of the historical dispersal
of invasive species. A map of the annual distribution of the Zebra mussel (Dreissena
polymorpha) is published by the US Geological Survey
4
and offers a dynamic view of
how species spread along waterways. The transformation of catchment areas by
anthropogenic activities are an important factor in increasing dispersal rates, such as
in the case the River Rhine (Leuven et al., 2009). In marine environments, the spread
of invasive species has also been evaluated and mapped, a good example being the
invasion of the Comb Jellyfish (Mnemiopsis leidyi) in the Caspian Sea
5
and the rest
of Europe.
The extent and speed at which the dispersal of species is achieved depends on the
inherent biological traits such as life history and reproductive strategy, as well as
environmental factors such as hydrodynamics and habitat suitability. Reviews on the
dispersal of a broad range of marine species such as presented in the work from
Bradbury et al. (2008) concluded an average dispersal rate of 39.8 km per generation.
Other works report the dispersal as an annual rate. Grosholz (1996) for example,
found an average dispersal rate of 50.7 km per year for 10 marine species, which is
very similar to what was reported by Byers et al. (2015) (37.8 km per year on average).
Lower estimates of dispersal rates may be reported for certain species, such as in the
case of the round goby, Neogobius melanostomus, where 25-30km seems to be a
better approximation (Steingraeber and Thiel, 2000, Azour et al., 2015).
To be correct, the rate of dispersion should be reported according to the life history
traits of the studied organisms. For the dispersal of the green algae, Caulerpa taxifolia,
which spread via fragmentation (Ceccherelli and Cinelli, 1999) or a coral for which
only the larvae is pelagic and therefore dispersed through currents, the reporting of
dispersal on a generation basis is warranted. For a holo-planktonic organism or a
migratory fish, the dispersal rate per year may seem more appropriate. However, the
lack of consistency in the reporting (temporal scale) may generate difficulties in
comparison exercises and one must look at these data cautiously.
3.5.2 Approach to evaluate the dispersal of organisms
For holo-planktonic organisms which live their entire life cycle in the water column,
the natural drift of all stages is to be taken into account in the evaluation, but for a
large amount of coastal marine species which are benthic and have limited adult
movement, the pelagic larval phase provides the primary mechanism of dispersal
(Cowen et al., 2006).
Spatio-temporal data on hydrodynamics can provide good insight as to the potential
risk of dispersal of key species. In the most basic form, hydrodynamic data describing
the dominant current regime and variation, tidal information, and seasonal variabilities
can be used to evaluate marine connectivity (Cowen et al., 2007). Such data may be
found published in papers, reports, books or tide tables. However, a successful
dispersal would rely on the movement mechanism driving the dispersal, the ability to
survive the dispersal, and the ability to mate, reproduce and produce successful
offspring at the new location after dispersal. For example, the European green crab
(Carcinus maenas) was first recorded in the US in 1817, followed by an episodic
expansion of range to the north. In eastern North America, it took almost 100 years
4
http://nas.er.usgs.gov/queries/SpeciesAnimatedMap.aspx?speciesID=5
5
http://www.grida.no/graphicslib/detail/comb-jelly-mnemiopsis-leidyi-spreading-through-the-caspian-sea-
invasive-species_795e and http://www.grida.no/graphicslib/detail/how-the-comb-jelly-mnemiopsis-leidyi-is-
spreading-through-european-seas-invasive-species_0524
MEPC 70/INF.21
Annex, page 20
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
for the green crab to expand its range from southern Massachusetts to the Gulf of
Maine. It was not recovered in northern Nova Scotia until the 1980s. Molecular data
suggest that this sudden and recent extension northwards is due to the recent
invasion of new genetic lineages with higher genetic diversity (Roman, 2006).
Indeed, most statistics on the dispersal of invasive species do not distinguish the
unaided natural dispersal from the (shipping) vector driven, and the numbers provided
on abundances are often the sum of both. New molecular data and genetic markers,
such as that applied in the example for the European green crab above, can be useful
in identifying genetic connections between populations and helping to ascertain the
likely sources, whether natural or introduced.
In such cases, modelling (biophysical, hydrodynamic (HD) and/or agent-based) can
be used and are robust approaches. The dispersal modelling is a comprehensive and
mechanistic approach that is able to simulate the dispersal of marine pelagic stages.
This can consist of two-dimensional (2D) or three-dimensional (3D) HD modelling,
larvae dispersal modelling (agent-based modelling (ABM) or particle tracking) and
connectivity matrix analyses. The HD modelling focuses on the motion of water and
simulates the dynamics of currents, waves and other hydrological features, which is
vital in the study of dispersal patterns and rates. Rather than solely focussing on
modelling the movement of water, ABM simulates the movement of individual organisms,
where each organism or autonomous "agent" moves according to a set of behavioural
rules and algorithms. In the example of pelagic larval dispersal, parameters may include
the pelagic larval duration, impact of passive drift and other behavioural mechanisms if
applicable, such as swimming in response to light. Examples of such ABM is detailed in
Section 5. Connectivity assessment can be performed by translating ABM result data
into 2D connectivity matrices, and such matrices represent probability of pairwise
connectivity (Treml et al., 2008, Jacobi et al., 2012).
As introduced earlier in this paragraph, the evaluation of natural dispersal is key to
determining the risk from ballast water mediated transfer. In this context, the modelling
developed to assess the natural spread of organisms may, when appropriate, take
into account the primary, secondary, tertiary etc. dispersal events over years and
generations.
4 Evidence for Connectivity in the inner EAS
With regards to delimitating a SRA, LMEs or smaller regions might need to be
considered. This flexibility is supported in the Guidelines (G7) (paragraph 6.2.3) which
stipulates: "It is recognized that the suggested biogeographical scheme may not be
appropriate in certain circumstances and in this case other recognized
biogeographical schemes may need to be considered."
As a result of comparing various biogeographic schemes (Section 3.3), the inner, non-
oceanic bioregions of the IMO Globallast EAS region (EAS-I to EAS-III and EAS-VI)
were chosen as a point of reference for this study, being focussed on Southeast Asia
(Figure 3.1). Furthermore, the IMO Globallast scheme is accepted among IMO
member states. The inner EAS region includes parts or all of the territorial waters of
Singapore, Malaysia, Indonesia, Brunei Darussalam, Thailand, Cambodia, Vietnam
and Philippines.
Scientific examples and methodologies describing marine connectivity (see
Section 3.4 for a definition) in this region with respect to the influence of hydrodynamic
and meteorological conditions, and dispersal characteristics of target species are
MEPC 70/INF.21
Annex, page 21
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
explored in the following paragraphs. However, it is important to keep in mind that the
present paper does not represent a SRA study but merely offers a list of examples
supporting the methodologies that may be applied in a final study defining a SRA for
the region. As such, the final boundaries of an eventual SRA in the region should be
verified through a detailed study.
In the following examples, the evidence for connectivity is explored across two
different spatial scales (regional seas level and local to the Singapore Straits),
recognizing that ecological connectivity in the sea has different characteristics at local
and regional scales relevant to scale-dependent parameters in hydrodynamic
processes and settlement processes.
Selected information's presented here below includes publicly available studies and
unpublished reports for the Singapore Government. The following information is used
as examples of methodologies applicable in the context of Same Risk Area:
Regional seas level:
- Cluster analysis of coral reefs and hydrodynamic modelling for the
South China Sea
- Flushing rates and residence times of water bodies in the Sunda
Shelf
- Particle-tracking analysis across the inner EAS region
Local to Singapore:
- Coral larvae dispersion
- Sea-star dispersion
4.1 Introduction to the EAS
This section provides some background to the marine characteristics of the EAS
region. The topography of the region includes shallow continental shelves such as the
Sunda Shelf, deep sea basins, troughs, trenches, continental slopes and volcanic and
coral islands. The numerous islands of various sizes divide the marine waters into
different seas connected by many channels, passages and straits (Kelleher, 1995).
The South China Sea is the largest sea in Southeast Asia resulting in a semi-enclosed
large marine ecosystem (UNEP, 2004a). Circulation in the South China Sea is
influenced by the twice-annual monsoons. The northeast monsoon towards the end
of the year forces surface currents north to south, while the south-west monsoon that
occurs mid-year drives currents in the reverse direction. There is also a flow of Pacific
water into the basin interacting with water from the Indian Ocean.
Temperatures and salinities vary spatially throughout the EAS region depending on
the inputs of oceanic water and river runoff and temporally depending on the monsoon
influences (Kelleher, 1995).
The Indo-West Pacific marine biogeographic area is a global hotspot of marine
tropical biodiversity (UNEP, 2004a). Fifty of the seventy global coral genera occur in
this marine area (Tomascik, 1997). Over 450 coral species have been recorded from
the Philippines compared with only 200 from the Red Sea, 117 from south-east India,
57 from the Persian Gulf and 35 from the Atlantic (UNEP, 2004a). Furthermore, over
half of Southeast Asia's hard coral species diversity is found in the South China Sea.
MEPC 70/INF.21
Annex, page 22
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
It is recognized that the location of the EAS at the junction between the Pacific and
Indian Ocean basins, has resulted in it becoming a centre of aggregation of species
from both Oceans. Kelleher (1995) noted a "basic homogeneity caused by the
occurrence of many wide-ranging species" within the EAS region.
The EAS region also supports a significant number of endemic species. For example,
only 5% of the 1,500 species of sponges in the South China Sea are distributed
throughout the Indo-West Pacific (Hooper et al., 2000), and 12% of the 982 species
of echinoderms in the South China Sea are endemic (Lane et al., 2001).
Another example supporting the rich biodiversity of the EAS region is also evident
among holoplanktonic organisms. In the Indo-Pacific area where Singapore is
situated, the encountered copepod biodiversity is one of the richest in the world (635
species) while other geographical areas of comparable size such as the North
Sea/English Channel/Norway/Baltic Sea area only counts 380 species and the coast
of California records 441 species
6
.
4.2 Regional connectivity
4.2.1 UNEP/GEF South China Sea Project
The UNEP/GEF South China Sea Project was a significant Global Environment
Facility project implemented by the United Nations Environment Programme from
2002 to 2008 to develop a regionally co-ordinated programme of action designed to
reverse environmental degradation particularly in the areas of habitat degradation and
loss, land-based pollution, and fisheries. Cambodia, China, Indonesia, Malaysia,
Philippines, Thailand, and Vietnam were participating countries in the study
(UNEP, 2009). A South China Sea Meta-database and GIS was developed enabling
open access to a large amount of technical reports and data.
As part of the project, specific activities were carried out that are relevant to this Same
Risk Area study, namely:
An assessment of the ecological status of mangroves, non-oceanic coral
reefs, seagrass and coastal wetlands; and
Establishment of a regional HD model to inform nutrient-loading and
carrying capacity models.
Short regional overviews of each of the four habitat types were published as inputs to
the first regional scientific conference (UNEP, 2004b, UNEP, 2004a, UNEP, 2004c,
UNEP, 2004d). A total of 136 habitat sites, including 26 mangrove, 43 coral reef, 26
seagrass and 41 coastal wetland sites, were characterized. A cluster analysis of coral
reef sites indicated some similarities among sites that may be used to infer ecological
connectivity, although levels of biodiversity may be influenced by a range of other
factors. The sites illustrating the highest levels of biodiversity were situated around
the edge of the South China Sea.
The land-based pollution component of the project aimed to evaluate carrying and
assimilation capacity of sub-regions and sensitive ecosystems and transboundary
movements of contaminants within the South China Sea.
6
http://copepodes.obs-banyuls.fr/en/
MEPC 70/INF.21
Annex, page 23
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Modelling of the assimilative capacity of the South China Sea marine basin with
respect to nutrient contaminant inputs from land-based sources was initiated resulting
in a simple Excel based model that can be downloaded from the project website
7
and
run on a desk-top, at basin or sub-basin scales to generate scenarios of potential
impacts consequent on changes in riverine inputs of nutrients to different sections of
the South China Sea coastline (UNEP, 2007).
The nutrient carrying capacity model developed by the study couples ocean
circulation with a biogeochemical mass balance model to simulate the geographic
distribution of phytoplankton biomass and anoxia in both surface and deep layers of
the South China Sea and under different scenarios of nutrient loading from adjacent
rivers. It is based on the Princeton Ocean Model (POM) circulation model run for the
defined domain using wind and density as the primary forcing functions. The reason
for this is that in an open water system where the horizontal gradients of contaminant
and particulate concentrations are usually small, mixing does not cause large net
horizontal transport and only the advection term is significant. This can be practically
estimated using an ocean circulation model such as the POM.
The model was run both in barotropic and baroclinic modes so that both depth-
integrated transport and velocities in surface and deep layers were obtained. The
upper or surface water was defined as the layer between the sea surface and a depth
of 50 m or the sea bottom, whichever was less. Surface circulation or current outputs
were verified against Japan Oceanographic Data Center (JODC) ship drift data.
MODIS satellite-based chlorophyll concentrations were used to calibrate the present
monthly nutrient loading from 190+ rivers and national data was used to verify such
loads for some rivers.
UNEP (2007) noted that over most of the South China Sea continental shelf, water
depths of approximately 50-80 metres generally coincide with the thickness of the
surface, mixed layer. This indication that the surface, mixed layer is vertically
homogeneous meant that modelling could be carried out on a 2-dimensional basis
simplifying both the modelling and data requirements substantially.
A study by Zhang and Su (2006) illustrated that the hydrographic properties of water
masses on each side of the Bashi Channel, the largest channel that connects the
South China Sea with the Pacific, were distinct over distances of only a few hundred
kilometres suggesting limited exchange between the two seas. Furthermore, most
channels/straits connecting the South China Sea with adjacent water bodies, such as
the Western Pacific Ocean, Sulu Sea and Java Sea are of narrow widths and shallow
depths. As a result the UNEP study assumed that the exchange of water and nutrients
between the South China Sea and adjacent water bodies was zero.
With respect to the HD model, UNEP (2007) illustrated a reasonable match between
the monthly aggregated surface velocity currents from the POM model and the ship
drift data (Figure 4.1). During the NE/NW monsoon, currents enter the north-eastern
part of the Sunda Shelf from southern China and through the Luzon Strait. They follow
the coastline of Vietnam southward along to the West-Malay coast. Here it partly
leaves through the Singapore Strait into the Malacca Strait and partly proceeds
through to the Karimata Strait to the south. A small eddy circulates in the Gulf of
Thailand and eddies appear to spin off from the main flow along the northern coastline
of Borneo.
7
http://www.unepscs.org/nutrient_model/
MEPC 70/INF.21
Annex, page 24
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.1 A comparison between JODC current implied from ship drift data (left) and POM
output in the UNEP (2007) study. Data were aggregated to 1-degree. Source:
UNEP (2007)
No correlation was found between observed surface chlorophyll biomass and the
oxygen concentration in the water layer immediately beneath the surface, mixed layer
in the South China Sea. The deep water in many parts of the South China Sea appeared
to be better oxygenated than predicted by the model, which might be due to a
re-oxygenation of the water mass through physical process in the deep water layer.
4.2.1.1 Conclusion for the SRA
The modelling exercises undertaken for the UNEP/GEF study provide a useful
reference to help refine further regional scale modelling. Although the 2-dimensional
modelling approach and parameters used helped to simplify both the modelling and
data requirements substantially, modelling of dispersion along coastal areas will
require greater resolution and the consideration of tidal forces.
Despite these limitations, the modelling illustrated the hydrodynamic flow throughout
the whole of the EAS-I during the NE/NW monsoon and indicates the potential for
ecological connectivity at a coarse level. It would appear that modelling to confirm the
boundaries for an SRA in this area should be extended to the whole of the inner EAS
in order to explore connectivity at the edges of a potential SRA and to answer the
question of whether a definitive boundary can be drawn. The coarse modelling
illustrated above does not indicate any obvious boundaries within the inner EAS with
respect to hydrodynamic patterns although eddies can act as boundaries to dispersion
by entraining larvae. This will be elaborated on in the ensuing case studies below.
4.2.2 Flushing rates and residence times of water bodies in the Sunda Shelf
A study on the exchange rates of water masses in the different parts of the Sunda
Shelf by Mayer et al. (2015), provides further evidence of the patterns of
hydrodynamic connectivity in the region. The Sunda Shelf is a shallow continental
shelf within EAS-I adjacent to southern Vietnam, eastern Thailand, Malaysia and
Singapore. This case study is described further below.
4.2.2.1 Methods
Four numerical models were used to estimate the flushing rates and residence times
in different seas on the Sunda Shelf:
MEPC 70/INF.21
Annex, page 25
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
The Max Planck Institute Ocean Model (MPI-OM) was used to simulate
the global ocean circulation. MPI-OM is the oceanic part of the climate
model system used for the Intergovernmental Panel on Climate Change
world's climate simulations. It had a horizontal resolution of 24 km
(approximately 44 km at lower latitudes) and 40 z layers of increasing
thicknesses from top to bottom.
The model above delivered the open boundary forcing data (vertical
structure of temperature and salinity, sea surface height) as six hourly
time series for the second component of the model system, the regional
ocean circulation model HAMSOM (Hamburg Shelf Ocean Model). It has
a horizontal grid resolution of 6 min (approximately 11 km) and 39 z
layers with increasing thicknesses from 5 m at the surface to 550 m at
greater depths.
Freshwater inflow at the river mouths were generated with the MPI-HM
(Max Planck Institute Hydrology Model).
Finally, a Lagrangian tracer model, based on a suspended particulate
matter transport model, was used to passively follow the water flow.
A single particle was placed into the centre of each model grid cell being
part of the region of interest. The trace of its path and the time needed
until it leaves the certain region is stored. Flushing rates were calculated
for each domain or area, standardized to the volume of the domain.
As with the rationale given for the UNEP/GEF study, only advection plays a significant
role in this case and turbulent mixing was considered to be irrelevant. The simulation
period covered the years 1958 to 2012. Model results were averaged over 24 h. Tidal
forcing was switched off but this was unproblematic as the study was focussed on
long-term phenomena and behaviour where the tides play a minor role. In order to
calculate the water renewal parameters, the regional model's horizontal velocities
were averaged to receive monthly values for each of the five decades from 1960
to 2012. Validation of the model results were made by comparisons with observed
velocities from moored current meters at different key locations within the Indonesian
Seas.
4.2.2.2 Case study findings
Most regions have similar flushing rates ranging from approximately 35 to 65 days
while the Gulf of Thailand is an exception being flushed at rates of 75 to 170 days.
The influence of seasonality was also greatest in the Gulf compared to the other
regions with lowest flushing rates in July and August and highest in May and October.
During the inter-monsoon phases, April/May and October, there is some stagnation
of the circulation, which changes its direction in some areas. This leads to a
deceleration of water currents and longer flushing rates. No decadal trend was visible
in the monthly results and inter-annual variability was not investigated because of the
general high variability of the current system in the Indonesian Seas (Mayer and
Damm, 2012).
The tracer model provides quite different but very detailed 3-D pictures (Figure 4.2)
with vertically averaged residence times of below 30 days to more than two years,
depending on the location within the region and on the season. An area of influence
from oceanic currents can be seen to the east of the Sunda Shelf as evidenced by
the blue areas of low residence time in Figure 4.2). This result does appear to support
the boundary for the eastern edge of the Sunda Shelf in the middle of the Bay.
MEPC 70/INF.21
Annex, page 26
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
However, the analysis needs to be extended to the northeast to examine the degree
of connectivity among coastal waters of the Sunda Shelf and northern Vietnam. Note
that these are monthly averages and it should be substantiated that they provide
correct results through finer scale temporal analysis.
Residence times also varied accordingly to the starting layer (Figure 4.3) with the
lowest residence times seen in the surface waters, 0-5m. The SE/SW monsoon
results in higher flushing of the Gulf of Thailand compared to the NE/NW monsoon
and inter-monsoon periods.
Figure 4.2 Vertically averaged residence times as simulated by the tracer model for the
Sunda Shelf from 2000 to 2012 based on monthly mean velocity fields for
February (NW/NE monsoon), August (SE/SW monsoon) and May and October
(Inter-monsoon) (Source, Mayer et al., 2015).
MEPC 70/INF.21
Annex, page 27
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.3 Residence times as simulated by the tracer model for three depth layers of the
Sunda Shelf region for 2000 to 2012 based on monthly mean velocity fields for
February (NW/NE monsoon season) and August (SE/SW monsoon) (Source,
Mayer et al., 2015)
4.2.2.3 Conclusions for the SRA
The study by Mayer et al. (2015) illustrates evidence for the flow of currents
throughout the Sunda Shelf region and the potential for hydrodynamic connectivity at
a coarse level. Furthermore, these results appear to support the gross boundaries of
an SRA based on the Sunda Shelf, at least for the eastern edge. However, this should
be confirmed by extending the HD model to the whole of the inner EAS in order to
explore connectivity at the edges of a potential SRA. The potential for connectivity
MEPC 70/INF.21
Annex, page 28
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
appears to be highest along the coastlines, therefore it would make sense to refine
the modelling in the nearshore and to explore local-scale levels of connectivity.
Analysis of residence times at multiple depths indicated that the majority of water
movement is in the upper 5m of the water column. This information can help to inform
later modelling used to define the boundaries of the study area for the SRA. However,
most of the dispersion appears along the coastal areas. Tidal forcing will need to be
considered in the model to better understand the finer resolution of dispersal patterns
at the coast.
The study by Mayer et al. (2015) indicated differences among different domains or
sub-areas of the Sunda Shelf with respect to flushing rates and residence times. In
particular, the Gulf of Thailand has on average the lowest flushing rates and highest
residence time of all areas in the Sunda Shelf. This indicates that under a SRA
approach, a multi-scale risk assessment would be advisable. For example, the Gulf
of Thailand with low flushing rates and high residence times will have a lower degree
of natural ecological connectivity to other ecoregions in the Sunda Shelf. It follows
that, given this higher degree of isolation, that there is a lower likelihood of potentially
invasive species naturally dispersing into the area from the surrounding ecoregions.
4.2.3 Particle-tracking analysis
Particle-tracking analysis was undertaken by DHI for this SRA study to simulate the
transport and fate of multiple suspended particles being released near the major ports
in the study area. The main objective of the case study has been to numerically model
the hydrodynamic conditions (water level and current conditions) of the inner EAS
region and the dispersion of discharged particles as a function of space and time so
that the potential dispersal path of particles could be identified.
4.2.3.1 Methods
Numerical modelling schemes for this study consist of DHI's 2D flow model MIKE 21
FM Hydrodynamic (HD) and Particle Tracking (PT) modules.
The modules are briefly introduced below:
The HD module simulates water level variations and flows in response
to a variety of forcing functions in oceans, lakes, estuaries, bay and
coastal regions. It is applied to a wide range of hydraulics related
phenomena, including tidal hydraulics, wind and wave generated
currents, storm surges and flood waves. Areas are flooded and dried
during a tidal cycle, both in nature and in the model.
The PT module is used for modelling the transport and fate of dissolved,
suspended and settled particles / substances discharged or accidentally
spilled in lakes, estuaries and coastal areas or in the open sea. With
respect to the SRA study, the release may be micro-algae or planktonic
larvae. The model assumes that the release behaves as particles being
advected by the surrounding currents and dispersed as a result of a
random 2D process. The particles may settle with a constant settling
velocity and settled particles may be re-suspended. A corresponding
mass is attached to each particle which may be reduced during the
simulation due to decay.
The model setups for both HD and PA modules can be found in the following section.
MEPC 70/INF.21
Annex, page 29
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Hydrodynamic model setup
A 2D regional HD model has been implemented to predict the water level and current
conditions in part of the inner EAS region and to capture the seasonal variations
induced by wind forcing or other effect. The model is still under development and
optimization but calibration using preliminary results have shown that its performance
is suitable for the present study to conceptually indicate the dispersal of particles that
are closely linked with hydrodynamic conditions. This is a high resolution model
(composed of small cells) that describes coastal areas in detail. It also includes the
description of short time scale processes such as tides that are not included in the
models previously described (Section 5.2.1.2 and 5.2.2).
The extent of the regional model has included the inner part of the EAS region not
influenced by oceanic conditions, i.e. EAS-I to EAS III and EAS-VI (Figure 4.4). Also
included to the west are part of the Central Indian Ocean (CIO) in order to explore
connectivity with EAS-VI and to the north-east is the North West Pacific (NWP) region
adjacent to EAS-I. There are seven open boundaries in the model domain. The
resolution of the regional model varies from 1 km at the Malacca Straits and the Riau
Archipelago to about 10 km near the open boundaries. The extent of the regional HD
model is shown in Figure 5.6, with open boundaries shown in yellow and the study
area shown in maroon.
The bathymetry information has been obtained from C-Map's electronic database for
the regional area. In some areas, this regional bathymetric data have been
supplemented by local survey data from other projects carried out by DHI in the
region. Depths from C-Map refer to Chart Datum (CD). All bathymetry data were then
converted to mean sea level (MSL) datum using a MSL map interpolated from mean
sea level values available from the C-Map database. For bed resistance, the manning
number, M is a function of the water depth and for the entire model domain, a manning
number map containing 32 to 45 m3/s has been applied.
The regional HD model is driven by time-varying water levels specified at the seven
open boundaries. For these boundaries, profile series of water levels have been
extracted from DHI's Global Tide Model with 0.125 degree resolution (DHI, 2014).
The 2D HD model was configured to apply barotropic mode in the flow calculation,
i.e. assuming that the density, the temperature and salinity are unchanged and will
not be updated during the simulation. The model could be otherwise configured to
apply baroclinic mode, for which the density is considered as a function of temperature
and/or salinity.
A 3D modelling will be useful to calculate the influence of the density variation over
the depth, particularly in deep sea where stratification due to temperature and/or
salinity is expected. The surface layer circulation in a 3D model may differ with that
from the depth averaged circulation. Moving forward to further develop the inner EAS
model, a three-dimensional flow model considering baroclinic computation of the flow
is necessary. It is highly preferred to carry out data collection in this region for model
calibration and verification. For the present DHI study, however, the 2D HD model is
suitable and useful to give first estimation of the processes in the study area.
Particle tracking model setup
Particles are divided into different groups called classes. Each class has specific
properties regarding decay, settling/buoyancy, erosion, and dispersion that has to be
specified separately. Typical examples of classes are different size fractions of
sediment particles, organic pollutants with different decay rates, floating objects or
coloured tracers.
MEPC 70/INF.21
Annex, page 30
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
For the present study, an artificial individual class has been setup without decay,
settling, and erosion so that the simulated path will be the longest path a particle could
possibly travel due to hydrodynamic condition and wind forcing throughout the
simulation period. The dispersion coefficient was assumed at 1 (which is the same as
eddy viscosity used in solution of the flow equations) in creating random movement
of the particles. A total of 11 releasing sources have been identified to represent major
ports in the study area, as shown in Figure 4.5. For each release source, a thousand
(1,000) particles were released once at the first time step (after model warm-up
period). The model setup has been established for conceptual illustration of the
relationship between hydrodynamic and particle tracking. It is recommended to revise
the modelling parameters for simulating the movement of a specific species or to
apply ABM in the later stage.
Figure 4.4 Extent and boundaries of regional model. Boundaries of the IMO Globallast
Regions and major seas are also indicated.
MEPC 70/INF.21
Annex, page 31
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.5 Locations of release points (1: Da Nang, Vietnam. 2: Ho Chi Minh, Vietnam. 3:
Laem Chabang, Thailand. 4: Mergui, Thailand. 5: Phuket, Thailand. 6: Port
Klang, Malaysia. 7: Tanjung Pelepas, Malaysia. 8: Singapore. 9:.Jakarta,
Indonesia. 10: Surabaya, Indonesia. 11: Brunei)
Climatic conditions
Climatic and meteorological conditions in the inner EAS region may differ throughout
the year. The particle tracking model, decoupled with the results from HD modelling,
was simulated for the three representative climate conditions, influencing tidal
conditions and seasonal weather patterns in the region, as described below.
North-east monsoon conditions (NE). Winds are predominantly blowing
from the NE direction. This season occurs typically from December to
February.
South-west monsoon conditions (SW). Winds are predominantly blowing
from SW direction. This season occurs typically from June to August.
Inter-monsoon conditions (IM). The duration between the NE and SW
are the inter-monsoon periods for which seasonal monsoon winds are
weakened.
Figure 4.6 shows the instantaneous wind magnitude and directions blowing over the
region for the three representative climatic conditions. The wind field has been
sourced from the National Centre for Environmental Prediction (NCEP), USA. The
data is based on the Climate Forecast System Reanalysis (CFSR) project using a
global, high resolution, coupled atmosphere-ocean-land surface-sea ice modelling
system. The data is available at hourly intervals, with a spatial resolution of 0.30°×0.30°
from 1979 to 2010 and 0.20°×0.20° from 2011 to 2013. CFSR wind data is
representative of 10 minutes average winds at a height of 10 m above mean sea level.
MEPC 70/INF.21
Annex, page 32
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
All climatic conditions are represented in 3.5-month simulation periods that cover the
entire monsoon seasons as well as neap and spring tidal cycles and 15-day warm-up
periods to avoid any type of instabilities that could occur during the initial state of the
simulations.
Figure 4.6 Instantaneous wind field map for NE (top left), SW (top right) and inter- (bottom
left) monsoon periods
4.2.3.2 Case study findings
The simulated water level is compared with the tidal levels (Table 4.1) at some
selected tidal stations in the study area (Figure 4.7) in terms of constituent amplitude
and phase. In general, the regional HD model has reasonably good agreement with
the tidal constituents in the area.
Table 4.1 Comparison of published tidal constituents and model simulation results. The
published tidal constituents are obtained from /1/ DHI KMS model
Station Name
Measurement
Model Simulation
Difference
Tidal
Constituents
Amp
(cm)
Phase
(deg)
Amp
(cm)
Phase
(deg)
Amp
(cm)
Phase
(deg)
Ko Miang
M2
67
86
78
84
11
-2
S2
32
112
34
115
2
3
K1
12
219
12
224
0
5
O1
5
194
5
189
0
-4
Iyu Kecil
M2
98
91
95
98
-3
8
S2
41
135
36
141
-4
6
K1
26
40
31
43
6
3
MEPC 70/INF.21
Annex, page 33
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Station Name
Measurement
Model Simulation
Difference
Tidal
Constituents
Amp
(cm)
Phase
(deg)
Amp
(cm)
Phase
(deg)
Amp
(cm)
Phase
(deg)
O1
23
6
20
11
-3
5
North Danger Reef
M2
18
76
25
81
8
5
S2
7
105
8
108
1
3
K1
32
191
34
195
2
3
O1
29
151
26
154
-3
3
Ile Bach Long Vi
M2
9
243
14
247
5
4
S2
4
282
4
277
1
-4
K1
65
336
61
330
-4
-6
O1
74
282
70
276
-5
-5
Ko Mak
M2
9
171
15
188
5
17
S2
5
232
6
238
1
6
K1
42
48
43
57
1
9
O1
29
8
25
11
-4
3
The results from the HD model were computed to give statistical mean u- and v-
components across the model domain over the 3-month monsoon simulation period.
The mean u- and v-components were converted to magnitude and direction, known
as the net current speed and direction to represent the residual current speed and
direction induced by wind. It assumes that the tidal induced currents are balanced
after averaging. Figure 4.8 to Figure 4.10 show the net current plot and density map
of time-accumulated particles during NE, SW and inter-monsoon respectively.
The modelling results have shown that the currents in the inner EAS is reflecting the
monsoonal wind patterns, that is southward in the NE monsoon and northward in the
SW monsoon. The monsoon-following currents are well-recognized (Guohong et al.,
2005, Wyrtki, 1961, Xu and Malanotte-Rizzoli, 2013).
Figure 4.7 Locations of tidal stations
MEPC 70/INF.21
Annex, page 34
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
During NE monsoon, net current flows along the near coast of Vietnam to the south
and splits in two directions at the southern tip of Vietnam with a portion of the flow into
the Gulf of Thailand and the main flow towards the Java Sea. Particles released in
this main flow path, such as off the coast of Vietnam and east coast of Peninsular
Malaysia will tend to flow south as they are driven by the hydrodynamic forces, as
shown by the net current direction. In contrast to the strong net current flow, particles
released in areas with relatively low net current speed, such as the Gulf of Thailand,
will tend to have weaker dispersal. For Singapore Straits, the net current is towards
Malacca Straits and the flow is towards the Andaman Sea.
During SW monsoon, the main net current flow is from the Java Sea to the South
China Sea. Particles released along the Vietnam coast and east coast of Peninsular
Malaysia tend to flow towards the South China Sea. At Jakarta, the particles tend to
flow into the Indian Ocean through the Singapore Straits. The Gulf of Thailand is
relatively calm and weak dispersal is expected. Net current in the Singapore Straits is
towards the South China Sea.
During inter-monsoon, the net current pattern does not show a strong predominant
direction due to the weakening monsoon wind conditions and the dispersal pattern is
between those shown in the NE and SW monsoons.
Eddies are illustrated in the Gulf of Thailand and around the Riau Islands in the middle
of the Sunda Shelf. Eddies themselves could indicate potential dispersal barriers as
they may entrain larvae, delaying dispersal and thereby causing mortality in species
with relatively short pelagic stages before they are able to reach a suitable habitat.
Note that the blue areas do not denote areas outside of the SRA. The particle-tracking
was limited to 3 months for simplicity for the purposes of this conceptual study. This
is the value of more detailed ABM over longer time periods of ten years that can
highlight the role of stepping stones and wider dispersal throughout the region over a
longer time scale. For example, in the first year a particle from northern Vietnam might
reach the southern coast and Gulf of Thailand. If this larvae settles and reproduces,
the following year it might make it down the coast to Singapore, and so on.
MEPC 70/INF.21
Annex, page 35
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.8 Net current (top) and particle path (bottom) during NE monsoon.
MEPC 70/INF.21
Annex, page 36
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.9 Net current (top) and particle path (bottom) during SW monsoon.
MEPC 70/INF.21
Annex, page 37
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.10 Net current (top) and particle path (bottom) during inter-monsoon
4.2.3.3 Conclusions for the SRA
A regional HD model has been set up to numerically model the hydrodynamic
conditions of non-oceanic, inner seas of the EAS region and the dispersion of
discharged particles during NE, SW and inter-monsoon to illustrate the potential
dispersal path of particles as driven by the hydrodynamic force.
The results of the HD model matches the coarse level results from the UNEP study
in illustrating the strong net current along the Vietnamese coastline. Areas with
MEPC 70/INF.21
Annex, page 38
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
stronger net current have shown stronger connectivity in linking the particles released
upstream of the main flow to the downstream, such as the Vietnam Coast. Areas with
weaker net current have shown lower connectivity with particles tending to disperse
locally, such as the Gulf of Thailand.
Though the present study provides some good estimates for the transport through the
inner EAS region, it is most likely that the estimates contain some uncertainty as 3D
effects and other local-scale processes can play a role on the transport across the
different regions. Therefore, it is considered relevant that for later stages of the project
further data collection, model development and calibration is carried out to allow for
more accurate predictions.
The model setup has been established for conceptual illustration of the relationship
between hydrodynamics and particle tracking. It is recommended to revise the
modelling parameters for simulating the movement of a specific species or to apply
ABM to confirm the boundaries of the SRA.
A 3D modelling is recommended to calculate the influence of the density variation
over depth, particularly in deep sea where stratification due to temperature and/or
salinity is expected. The surface layer circulation in a 3D model may differ with that
from the depth averaged circulation. Moving forward to further develop the inner EAS
model, a three-dimensional flow model considering baroclinic computation of the flow
is necessary. It is highly preferred to carry out data collection in this region for model
calibration and verification.
4.3 Local -scale connectivity
Regional-scale results above indicate the need for detailed modelling of processes in
the near-shore and at a local scale. The following two examples illustrate local
connectivity processes among Singapore, Malaysia and Indonesia. In particular, they
illustrate the influence of small islands and peninsulas in determining larval dispersal
rates and source-sink patterns.
4.4 Conclusions for the EAS Region
In summary, the evidence from a range of studies at different levels of resolution
indicate hydrodynamic connectivity throughout in the EAS region at both regional and
local scales. Seasonal patterns are evident with the highest flushing rates during the
alternating monsoons resulting in reversals of current flow at regional and local scales
and bi-directional connectivity. The potential for connectivity appears to be highest
along the coastlines.
The boundary of the Sunda Shelf, particularly to the west and south, was largely
supported by the coarse resolution modelling by Mayer et al (2015) and the
preliminary particle tracking analysis carried out as part of the present study at a finer
resolution by DHI. Both studies indicate the potential for connectivity among coastal
waters of southern Vietnam and the Gulf of Tonkin to the north. Studies to confirm the
boundaries of a Same Risk Area should be extended beyond the EAS-I.
Spatial variations in connectivity and flushing rates indicate that a multi-scale analysis
of connectivity for the EAS region would be advisable. For example, the Gulf of
Thailand with low flushing rates and high residence times is expected to have a low
degree of natural ecological connectivity to other parts of the EAS. A multi-scale
analysis can be achieved through the nested boundaries of the IMO Globallast
MEPC 70/INF.21
Annex, page 39
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
regions and seas (Kelleher et al. (1995)) with some further nesting to capture finer
local scales as illustrated in the coral and sea start dispersion studies.
For most of the EAS-I, modelling can be carried out on a 2-dimensional basis as the
surface, mixed layer is vertically homogeneous in most areas due to the shallow water
depth characteristics of this water body. The results from the 2D modelling exercise
that include local-scale processes illustrate the role that small scale (time and space)
hydrodynamic tidal patterns and geographical barriers such as islands can play in
determining larval dispersal and successful settlement. This can be observed, for
example, in high tidal flows areas along the Singapore Straits where flows run parallel
with the border between Singapore and Indonesia and during monsoon periods may
serve as a barrier to dispersal across the two land masses.
However, on the edge of the continental shelf and in deep water the stratification and
wind effects are important and the 2D model may under predict the surface currents
in these areas. Based on the flow modelling results there is a strong argument for the
inclusion of higher resolution 3D modelling to describe both deep ocean effects and
nearshore tidal forces to fully describe the patterns of dispersal. It is further important
to choose parameters carefully so that the patterns of interest are not missed or
disappear into seasonal averages.
More generally, these studies show how modelling tools can be used to illustrate the
scientific evidence for marine connectivity and dispersal characteristics of marine
larvae. The use of modelling tools is therefore very useful to support the evaluation of
bio-invasion scenario.
4.4.1 Coral larvae dispersion
An investigation of coral larvae recruitment potential (including the connectivity and
dispersal patterns of coral larvae) at the Sisters' Islands and Labrador Nature Reserve
of Singapore was carried out for the National Parks Board of Singapore (NParks) to
assess the suitability of the sites for further enrichment of the coral habitat.
Different coral species have various strategies for reproduction and dispersal,
however, a large proportion of species are known as broadcast spawners; these
corals release vast quantities of gametes and sperm directly into the water column for
fertilisation to occur (Lipcius et al., 2008). These bundles of positively buoyant
gametes and sperm float to the surface where their dispersal, given their limited
mobility, are almost entirely dependent on the water currents (Black, 1993).
4.4.1.1 Methods
Timing
As the dispersal of broadcasting coral in its initial life stage is predominately driven by
ocean currents and winds, the timing of the spawning event plays a key role in the
resulting distribution patterns. The timing of spawning of broadcasting coral species
in Singapore is well documented through an extensive DHI coral reef monitoring
programme, and usually occur over a period of 3-5 days following a full moon in late
March/beginning of April. As it is generally estimated that the April mass spawning
events account for about 50% of the total amount of spawned larvae over the course
of a year, this study only considered the dispersal and connectivity patterns arising
from the observed mass-spawning event that took place from 18 to 20 April 2014
under the assumption that this event would be sufficient to explain the majority of
connectivity patterns in Singapore.
MEPC 70/INF.21
Annex, page 40
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Furthermore, while the maximum lifespan of coral larvae can reach >100 days
(Graham et al., 2008), it is assumed that the vast majority of larvae will either settle
or perish within the first ~25 days from spawning, and thus the chosen model only
covered the period from 16 April to 1 June 2014.
Identification of reefs
Based on the available GIS data-set of reef extent in Singapore, individual reefs were
divided into thirty-five geographically distinct reef groups (Figure 4.11). Each of the
reef groups act as an individual larvae source and sink in each model scenario
(35 individual simulations), in order to establish the variation in dispersal and
settlement patterns due to the geographic location of the reef. Substrate quality of
each reef group was calculated based on the ratio between the recorded reef area
within each reef group and the area of model element cells used to resolve each reef
group. Furthermore, only 10% of the total reef area was assumed to be suitable for
settlement.
Figure 4.11 Model included reefs within Singapore (above) and overlaid with the
computational mesh (below); note high horizontal model resolution around all
included reefs
Model description
In ABM Lab, the actual movement of agents in the horizontal plane are expressed by
assigning values (constant or variable over time) as a horizontal vector, consisting of
MEPC 70/INF.21
Annex, page 41
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
a speed (m/s) and a direction (degrees). Movement in the vertical plane is defined
through a negative (swimming upwards) or positive (swimming downwards) speed. At
any given time the simulated agents receive information of their surroundings in their
current location on the Eulerian grid, e.g. agents can sense their immediate
surroundings and be made to react to it through user-defined expressions.
Transport by currents or passive drift was calculated using a combination of advection
and dispersion, the former determined by the current velocity predicted by the HD
model, and the latter calculated using a random walk method and representing the
movement of agents in response to currents not resolved by the HD model, at a scale
smaller than the applied model resolution. The resulting transport of individual agents
was derived from a simple numerical solution to the Langevian equation which is
typically applied for simulating advection-dispersion processes.
Once fertilized, zygotes will start to develop into competent larvae, capable of settling
onto suitable substrate. However, due to significant predation rates as well as very
low natural survival rates (Graham et al., 2008), settlement success relative to the
initial number of spawned gametes is expected to be diminishingly small.
While a wide array of other environmental conditions affect the dispersal, survival,
and settlement of coral larvae, from a modelling perspective, it is necessary to make
a range of simplifying assumptions in regards to describing the governing forces that
control distribution patterns of coral larvae as follows:
The daily mortality rate, accounting for both natural mortality and death due
to predation and other environmental effects are expressed in the model by
a Weibull function, mimicking a type III survival curve, adapted from
(Graham et al., 2008).
The probability of gaining competency, e.g. progressing from a zygote
stage to competent larvae, is a function of the minimum incubation time
before a zygote can develop competency, and the maximum daily
probability of gaining competency and the age in days where a zygote
experiences the maximum daily probability of gaining competency.
The daily probability for competency loss as a function of the maximum
daily probability of larvae losing competency, and the time in days before
coral larvae experience the highest probability for losing competency.
The probability of settlement success in the model is directly related to
the substrate quality index ranging in values from 0-1, where 1 is
considered optimal habitat and zero is not optimal.
In order to settle successfully on substrate, each individual particle has
to pass some conditional statements: the particle has to have gained
competency, particle x,y-coordinates in the given model time step must
coincide with a reef area, particle needs to be less than 0.5 m away from
reef substrate (vertically) and current velocity needs to be below 0.5 m/s
in the moment of settlement. If these all hold true, then the settlement
chance equal the value of the substrate, e.g. a reef with an optimal value of
1, will equal a 100% settlement success chance.
Figure 4.12 provides a simplified flow diagram of the applied coral larvae dispersal
model, which depicts the various conditional statements that are checked for each
model time-step.
MEPC 70/INF.21
Annex, page 42
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.12 Simplified illustration of the main conditional expressions controlling the life,
death and settlement success of individual agents for each model time-step;
Note: The only exit to the loop of the flow-diagram is to either successfully settle
or perish
Model Parameterization
All simulation runs were set up with the same values of model parameters. This was
done in order to solely investigate the effect of the source location, substrate quality
and the effect of hydrodynamics between locations, rather than the effect of different
mortality rates or competency gain.
It is important to note that there exists a relatively high degree of natural variation and
uncertainty in each model constant relating to the biological properties of simulated
zygotes/larvae. However, as long as the applied values are kept constant between
MEPC 70/INF.21
Annex, page 43
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
simulations, the bias that is inherently associated with the value of a given constant
will be identical between simulations, and therefore not the cause of any potential
difference observed.
Sensitivity Analysis
In HD models, the horizontal and vertical dispersion usually describes transport due
to non-resolved processes in the model. In coastal areas, it can be transport due to
non-resolved turbulence or eddies. Especially in the horizontal direction, the effects
of non-resolved processes can be significant, and it is therefore usually one of the key
calibration parameters for dispersal studies.
The sensitivity of the model to horizontal and vertical dispersion was investigated in
depth in a previous study by DHI (DHI 2013, DHI 2014) and in general resulted in a
standard deviation of ~50% of the predicted mean settlement success, but with only
relatively minor deviations in emergent connectivity patterns. Instead of re-running the
same sensitivity analysis for this study, the magnitude of horizontal & vertical
dispersion were parameterized with the values that were previously found to generate
values close to the statistical means arising from the sensitivity analysis.
Model Calibration & Validation
Unlike hydrodynamic modelling, for ABM modelling proper calibration data
(e.g. concentrations of larvae over time, or number of new recruits) is rarely available.
Therefore, neither calibration nor validation of the coral larvae ABM was considered
in a traditional sense. However, a qualitative approach to calibration was taken in
order to ensure that the model obtained patterns that were deemed plausible relative
to what is known about the organism/system. For example, the model's ability to
predict known recruitment at sites such as Keppel Bay, Tuas View Extension and
Sultan Shoal was tested. Furthermore, the model was calibrated such that the
percentage of settled larvae from the vast majority of source reefs were in the range
of 0.5 to 1.5% relative to the amount of released agents per simulation.
Post-Processing Methods
Model results of simulated larvae tracks and resultant settlement locations, were post-
processed in order to obtain the following parameters:
Settlement Success: Defined as the number of larvae relative to the total
number of larvae released for each model scenario, which successfully
settled on reef substrate. Settlement success rates were scaled relative
to the total area of the reef group of origin, in order to account for
variation in spawning potential assuming a linear relationship with total
reef area.
Source Rank: Rank value assigned to each source reef relative to their
weighted settlement success and the number of reef groups that they
seed to.
Sink Rank: Rank value assigned to each reef group relative to the
number of reefs it receives larvae from, as well as the total weighted
amount of larvae it receives from other reefs scaled to the area of the
reefs own available substrate.
Source Connectivity: Defined as the percentage contribution of the total
amount of settled larvae to neighbouring reefs (sinks) from individual
source reefs. This measure helps establish the importance of individual
reefs as a source of larvae to other reefs.
MEPC 70/INF.21
Annex, page 44
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Sink Connectivity: Defined at the area-corrected contribution from
various source reefs to an individual reef. This measure helps assess
the potential resilience and potential genetic diversity of larvae received
for each reef relative to its primary sources.
Normalized Settled Larvae Densities: Maps of normalized settled larvae
per square meter were calculated for each individual reef in order to
visualize zones of high recruitment within the reef group. Due to the high
uncertainty in the true number of spawned larvae in nature, the absolute
values of larvae/m2 have been converted to normalized values ranging
from 0-100, 100 indicating the highest level of recruitment within the
spatial context of a single reef group.
4.4.1.2 Case study findings
Figure 4.13 illustrates the effects of larval dispersal from all coral reef sites in
Singapore 45 days after the "spawning event". The cumulative density of competent
larvae across to Indonesia to the south and Malaysia to the east and west can be
seen. Furthermore, the effect of the strong tidal currents in the Singapore Straits is
illustrated in the lower cumulative larval density in Indonesia, i.e. the majority of larvae
are transported east towards Malaysia.
The model outputs include source and sink rankings reflecting the number of reefs
seeded and the settlement success. Results revealed a relative uniform rank
classification of reefs located in the central parts of Singapore (e.g. Labrador and
Sisters' Islands reefs), with source and sink rankings generally varying between 6
and 7, out of 10.
There was a high number of contributing reefs to both Sisters' Islands and Labrador,
indicating a high level of connectivity and reasonably inferring a potentially high level
of genetic diversity and exchange. Both received larvae from relatively well distributed
reefs, which can be translated that both sites are predicted to be largely independent
from single source reefs. Both Labrador Nature Reserve and Sisters' Islands are also
good source reefs contributing to other reefs in Singapore to a reasonable degree.
These findings suggest that the reefs may be well-suited for habitat enrichment.
MEPC 70/INF.21
Annex, page 45
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.13 Temporal model snapshot of the cumulative density of competent larvae
dispersed from all coral reef sites in Singapore 45 days after the spawning event
On a regional scale, it is plausible that Sisters' Islands are more likely to receive larvae
from non-Singapore sources due to the proximity to the main Singapore Strait;
however, further studies will be required in order to define the importance of regional
connectivity patterns.
Semi-enclosed and geographically remote reefs all scored the lowest as source reefs.
The lowest rank in general could be attributed to high current velocities due to
proximity to the main Singapore Strait where currents reach ~2 m/s.
4.4.1.3 Conclusions for the SRA
With respect to the Same Risk Area, the results of this theoretical modelling exercise
reveal some of the characteristics of local-scale connectivity within the EAS region. In
particular, the case study reveals the role and location of source and sink reefs that
are important in the context of the stepping stone principles of dispersal. While
regional scale hydrodynamics may indicate larval connectivity between ports, the
successful recruitment of species at a local scale is highly dependent on other factors
such as the availability of suitable substrate and local current conditions.
This study shows how the modelling tools used by DHI in this study can illustrate the
scientific evidence for marine connectivity and dispersal characteristics of marine
larvae therefore supporting risk assessment studies.
4.4.2 Sea-Star dispersion
Between August 2012 and February 2014, DHI carried out a study to examine the
population dynamics of the echinoderm, Protoreaster nodosus, in Singapore, with an
aim to complement and add to the existing knowledge of this giant knobbly sea star
species. Specifically, the study provides key information on the spatial and temporal
patterns of abundance and size structure in local populations and the underlying
MEPC 70/INF.21
Annex, page 46
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
environmental conditions that influence the population dynamics. The purpose of the
study was to support management plans related to this species and possibly similar
seagrass-associated fauna. Of relevance to the study of Same Risk Areas are the
outputs from the genetic connectivity and larval dispersion modelling tasks.
In addition, monitoring studies indicated that spawning periods coincided with
favourable conditions in the gulf and individuals were observed to avoid intertidal
habitats possibly as a way to optimize the distribution of gametes.
4.4.2.1 Genetic connectivity of P. nodosus
Studies to investigate P. nodosus populations across 5 different locations located to
the north (Beting Bronok, Chek Jawa and Pulau Sekudu) and south (Changi and
Terumbu Pandan) off mainland Singapore using genetic techniques were conducted
starting in January 2013 (Figure 4.14). These techniques employing the use of genetic
markers are useful in providing evidence of populations that are closed, i.e. self-
recruiting, as genetic differentiation is highly sensitive to migration (Hellberg et al.,
2002). In Singapore, (Chim and Tan, 2013, Chim and Tan, 2012) successfully
described methods to tag and identify individuals. A total of 80 tissue samples were
collected from mostly adult individuals (one juvenile) at five localities in Singapore. Of
the 80 tissue samples collected, 78 were successfully sequenced for an extension of
the widely-used COI marker (COI-tRNA; (Crandall et al., 2008)). Sequences were
checked for quality, aligned, and haplotype diversity of the sampled P. nodosus
populations was assessed using 710bp of sequence data.
The study at the 5 locations indicated that the P. nodosus populations were
interconnected. Of the sites sampled, Terumbu Pandan, with the highest numbers of
individuals (approximately 900 counted during site selection surveys), is an area with
greater resilience to environmental disturbances due to its higher genetic diversity
and large population size. There were no clear source-sink relationships between the
selected sites and a possible indication of self-recruitment amongst one of the
populations in Singapore. These results indicate that more sites may have to be
investigated across a broader geographic area, with known locality records of P.
nodosus.
Numerical modelling studies to examine the larval connectivity of P. nodosus could
complement the genetic work outlined previously. Studies integrating ecological and
genetic data to investigate population connectivity have been carried out on other
ecological groups such as fish (Rivera et al., 2010), gastropods (White, 2010), and
polychaetes (Jolly et al., 2009). Modelling studies may complement work done for
marine species with weak population genetic structure. Nonetheless, quantifying the
''effective'' connectivity of marine populations may require validation using population
genetic tools.
MEPC 70/INF.21
Annex, page 47
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.14 Locations where sea star genetic tissue samples were collected
4.4.2.2 Larvae dispersal modelling
With the purpose of investigating the potential connectivity of P. nodosus within known
seagrass habitats in the coastal waters of Singapore, larval dispersal and connectivity
of selected seagrass beds were assessed in a preliminary study. To accomplish this,
an ABM of P. nodosus larvae was developed and implemented through the use of the
integrated MIKE by DHI's software module, ABM Lab.
The coupled Eulerian-Lagrangian framework in ABM Lab allows for an accurate
representation of hydraulics and water quality within a spatially complex system over
time, while it is also capable of simulating entities such as the dispersal and settlement
of planktonic larvae on an individual level. As simulated agents within the model
domain can be made capable of reacting to Eulerian gradients such as water
temperature or flow velocities, it is thus possible to investigate potential effects of
hydraulic and environmental cues on movement of aquatic organisms on a complex
spatial scale over time.
Hydrodynamic Model
To provide high resolution modelling results required around the smaller islands of
Singapore as well as computational efficiency, a previously DHI calibrated and
validated depth-averaged flexible mesh HD model for Singapore (MIKE 21 FM) was
used. The existing MIKE 21 FM was chosen due to its higher computational efficiency
and its capability of running the ABM module in a decoupled mode, where the agent-
based model can be run by reading the stored hydrodynamic parameters rather than
re-calculating the flow model for every single time-step. This significantly reduces the
computational time compared to classic structured mesh models.
Since seagrass beds located to the east and west were of particular interest to this
project, it was decided to increase the spatial resolution of the flexible mesh of the
original model to match that of the central Singapore inner harbour area (75m).
MEPC 70/INF.21
Annex, page 48
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Timing
P. nodosus are known broadcast spawners i.e. this type of species release vast
quantities of gametes and sperm directly into the water column for fertilisation to occur
(Yamaguchi, 1973). Bundles of geo-positive gametes and sperm will be released into
the water column where their dispersal, given their limited mobility, will be almost
entirely dependent on the water currents.
As the dispersal of P. nodosus in its initial life stage is predominately governed by
ocean currents, the timing of the spawning event plays a key role in the resulting
distribution patterns. While there is natural variation in the timing of spawn events, the
timing of spawning of P. nodosus in Singapore was assumed to take place in the
afternoon 3 days after the April full moon (Tun, pers comm.). Thus, the 2D FM model
was setup to simulate the period from 26th April to 7th June, 2013, with a model time-
step of 5 minutes, which was selected based on a compromise between
computational availability, while still being able to account for processes happening
on short time scales to a certain degree.
Identification of sources and sinks
For the purpose of assessing potential connectivity patterns between known habitats
of P. nodosus in the coastal waters of Singapore, four seagrass beds located at
different parts of Singapore: Terumbu Pandan, Pulau Sekudu, Pulau Semakau and
Beting Bronok were chosen as spawning sources, while all other habitats of interest
within the model domain acts as sinks on which larva can settle if the model settlement
criteria are met.
As a vital part of the model setup process, substrate suitability maps to be used for
model inputs were generated using GIS data as well as the information gathered from
DHI monitoring surveys, respectively. All habitat substrate were assumed to be
optimal for settlement and given the value of 1.0 (100% settlement success if other
model criteria are met).
Model Description and Parameterization
The model used was very similar to that previously described for coral larvae (see
Section 4.4.1.1). For example, a range of simplifying assumptions were made in
regards to describing the governing forces that control distribution patterns of P.
nodosus larvae (dispersal, survival, settlement, etc.). All simulation runs were set up
with the same values of model calibration parameters, with the exception of the
particle to zygote number being different between scenarios.
In 2D models, the dispersion usually describes transport due to non-resolved
processes. In coastal areas, it can be transport due to non-resolved turbulence or
eddies. Especially in the horizontal direction, the effects of non-resolved processes
can be significant, and it is therefore usually one of the key calibration parameters for
dispersal studies. Since the implemented HD model for this project applies a
Smagorinsky approximation to resolve the eddy viscosity, the horizontal dispersal of
agents (or larvae) are described using a scaled eddy viscosity formulation with a value
of the scaling factor ranging from 0.5-1.5.
Since the nature of a 2D model does not allow for the same kind of scaling of vertical
dispersion to eddy viscosity, a vertical dispersion coefficient with a value range from
0.001 to 0.01 (m2/s) was implemented instead. Aforementioned value ranges of
horizontal and vertical dispersion resulted in the following model calibration matrix.
MEPC 70/INF.21
Annex, page 49
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Case Study Findings
The inter-connectivity of the geographically separated populations was supported by
the larval dispersion modelling.
The initial dispersal of P. nodosus from Terumbu Pandan was very affected by the
prevailing westerly tidal currents at the time of spawning, which resulted in a sizeable
portion of zygotes being advected into the Tuas harbour area to the west (Figure
4.15), and thus becoming fragmented from the main plume. Due to the decreased
level of flushing inside the harbour basin, this fragmented population of larvae largely
remained within the harbour basin where it slowly perished due to lack of suitable
substrate for settlement. Note that the larval plume reaches the islands of Indonesia
to the south.
Figure 4.15 Temporal model snapshot of the initial dispersal of P. nodosus, from Terumbu
Pandan reef ~21 hours after the spawn event
The larval plume from Pulau Sekudu remained within the Strait of Johor for a
prolonged period before dispersing westwards and north into Malaysian waters
(Figure 4.16). The settlement success of the Pulau Sekudu larvae population was
significantly higher than the Terumbu Pandan population, with an overall mean
settlement success of 3.30% (± 0.37 SD) compared to 0.176% (± 0.037 SD) for
Terumbu Pandan. While settlement success was high, the mean displacement
lengths of settled larvae were only 2.16 km (± 2.72 SD), in contrast to 19.68 km (±
13.13 SD) for the Terumbu Pandan population.
MEPC 70/INF.21
Annex, page 50
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Figure 4.16 Temporal model snapshot of the initial dispersal of P. nodosus, from Pulau
Sekudu ~21 hours after the spawn event
A similar dispersion was illustrated for the Beting Bronok population with the dispersal
of larvae primarily limited to the extent of the eastern-most part of Johor Strait. The
initial dispersal of the Pulau Semakau population was affected by the seagrass beds
being sheltered by islands during the westerly currents at the moment of spawning.
Once dispersed into the main Singapore Strait, the plume kept being transported in a
predominantly east to west direction along with the direction of the tide, while
becoming more and more fragmented. The overall mean settlement success for Pulau
Semakau was a mere 0.156% (± 0.04 SD) which is slightly lower than the Terumbu
Pandan mean settlement success of 0.176%.
These results support the hypothesis that the geographically separated populations
of P. nodosus are genetically related through larval dispersal dynamics. The high
recruitment levels at Tanah Merah (relative to other locations) of ~3.5 larvae/m2 is
likely an effect of the oscillating motion of the tidal dynamics in the main Singapore
Strait, meaning that larvae dispersed into the main strait has a higher likelihood of
re-encountering habitat within the main strait area. Similar oscillating current patterns
are seen for the EAS Region (see Sections 4.2.2.2 and 4.2.3.2), albeit over larger
scales and longer seasonal time periods.
4.4.2.3 Conclusions for the SRA
The study illustrates that the population of P. nodosus in Singapore is unlikely to be
isolated from cross border habitats with dispersal across to Malaysia and Indonesia
indicated by the results. A risk assessment might have to consider a wider spatial
distribution of sites across a broader geographic area than Singapore in order to better
understand and confirm the risk of potentially invasive species translocating in and
out of Singapore.
The preliminary model results can also be further developed and strengthened
through the inclusion of a comprehensive data parameterisation e.g. substrate quality,
mortality rates, competency gain probabilities, as well as introducing smaller
time-steps and using a finer mesh resolution.
The study indicated a range of small-scale factors influencing local recruitment,
including the influence of geographical barriers such as islands and peninsulas. It is
recommended that the application of a SRA approach should be implemented at a
MEPC 70/INF.21
Annex, page 51
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
number of different spatial scales to understand how local factors interact with ocean-
scale hydrodynamics to influence the connectivity of species populations and
settlement success.
5 General Methodological Approach to Assessing a SRA
5.1 Draft Procedures
The workflow for an SRA approach to exemption is illustrated in Figure 5.1 and follows
a series of seven steps. These are the similar to those presented in MEPC 69/INF.25
with more detail and description added and the inclusion of mitigation measures under
Step Four.
Step One: Two or more States agree to investigate the possible use of SRA in a water
body shared by the States. A hydrodynamic (HD) modelling exercise would be
undertaken to define the preliminary zone of study for the SRA. An initial boundary for
the SRA could be based on the IMO Globallast Bioregions and subzones. Ports that
are only donors or only receptors would be removed from the SRA study. This would
help to inform which States to involve in consultation on the SRA.
Step Two: States agree on the species of concern, i.e. the target species. Each state
collates a list of target species which can be compared. If the same species is present
in all the selected ports, dependent on population characteristics, it could be omitted
from the assessment. If ports have different assemblages of target species they might
still be considered as part of the SRA, so long as they share the same environmental
conditions and are ecologically connected. This is based on an assumption that
although a target species has not yet made it from one port to another, it is just a
matter of time due to evidence on natural dispersal in the region.
Step Three: Given the target species, refined, multi-scale HD modelling and Agent
Based Modelling (ABM) is undertaken to assess their dispersal potential based on
their biological characteristics and the environmental conditions, e.g. salinity and
temperature. Relevant parameters used in the hydrographic modelling assessment
should also be agreed in advance. This might include the duration of dispersal and
the number of months/years (relative to generational time spans) over which to
average hydrographic statistics. The outcomes from these modelling exercises are
then used to determine the boundaries for the SRA. Independent scientific review of
the results is recommended at this Step to allow States to assess the validity of the
results in Step Four.
Step Four: Following the outcomes of the modelling and consultation exercises with
neighbouring States that may be affected, the extent of the SRA is agreed among
SRA States along with mitigation measures.
Step Five: Vessels trading in the SRA may apply for exemption under the terms of
regulation A-4.
Step Six: The outcome of the application (granted or rejected) is communicated to
the other SRA states and to the IMO.
Step Seven: Each vessel is issued the exemption to be recorded in the ballast water
record book.
MEPC 70/INF.21
Annex, page 52
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
The ensuing sections provide more detail on the scientific methodology and
considerations under Steps Two and Three.
Figure 5.1 Workflow of procedures required to conduct a SRA approach for exemption
5.2 Define Target Species for SRA
The G7 species-specific risk assessment description supports the use of selected
target species that are identified to have the ability to survive and reproduce in an
environment where they are not present and eventually becoming invasive. To carry
out such a risk assessment, the presence or absence description of species in donors
and recipient ports is necessary. Similarly, the SRA approach needs to identify a list
of target species to be used in the study.
There is a relative gap in knowledge with regards to invasive species in Southeast
and East Asia and in particular within the waters of the Sunda Shelf (Chavanich et al.,
2010; Molnar et al., 2008; Jaafar et al., 2012). This knowledge gap is partly due to the
lack of historical reports and a very high recognized biodiversity.
Therefore, in order to address this knowledge gap, Singapore initiated a number of
studies aimed at assessing the diversity of organisms in its territorial waters (i.e.
Schmoker et al. (2014) and others), and evaluating the status of ballast-mediated
marine species introductions in Singapore (Nparks, 2012_Unpublished). Additional
reports are being produced elsewhere in the region (see e.g. Kassim et al., 2015)
MEPC68.INF36).
5.2.1 Methods
Since one of the best predictors of the likelihood of a species becoming invasive is a
previous history of successful introduction in other habitat (Ricciardi, 2003), a list of
known invasive aquatic species was compiled by Nparks (2012_Unpublished) and
was used as a starting point for the analyses. A lengthy process of information
MEPC 70/INF.21
Annex, page 53
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
gathering was then initiated to define a list of target species as described in the step
by step process flowchart (Figure 5.2).
All species names and taxonomic groupings were checked against WoRMS (World
Register of Marine Species) to reduce the redundancies in species taxonomic names
and synonyms but entries that did not describe species names were omitted from the
study. Only species which were considered transportable by ballast water were used
(species which have been reported in ballast tanks or most likely to be found in the
water column during ballast uptake). In the case of meroplankton, the species were
only kept as part of the study if they had been reported as transported by ballast water.
Using environmental-matching risk assessment to identify target species as defined
in G7 Guideline, the study focused on invasive species that were recorded in tropical
zones, which are environmentally similar to Singapore. Invasive species that are
typically found only in temperate or polar latitudinal zones were therefore excluded.
The search was further refined to shortlist only invasive species that were present in
the Sunda Shelf because ships in trans-oceanic shipping are not likely to apply for
exemptions. Finally, the short-listed species were separated into two groups:
Blacklist: Ballast-mediated aquatic invasive species reported to be
present in Singapore
Watchlist: Ballast-mediated aquatic invasive species not currently
reported to be present in Singapore
Figure 5.2 Flowchart of data mining process relevant to create a blacklist and watch list
(Source: NParks, 2012 unpublished)
To determine the risk level of watch-listed species establishing in Singapore waters,
a preliminary qualitative risk assessment was carried out based on the temperature
and salinity tolerances of the target species and the environmental parameters known
to occur in the east Johor Straits which was taken as 27.6-32.3°C and 19-33PSU
(Gin et al., 2006).
The risk assessment matrix comprised three levels of risk:
MEPC 70/INF.21
Annex, page 54
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
Low risk: if neither parameter—temperature range nor salinity range—
matches that in Singapore's marine environment, the species is
categorized as low risk
Medium risk: if only one parameter—either temperature range or salinity
range—matches that in Singapore's marine environment, the species is
categorized as medium risk
High risk: if both parameters—temperature range and salinity range—
match that in Singapore's marine environment, the species is regarded
as having high risk of establishment.
In addition to the risk level generated for each species, the presence/absence of any
negative ecological, economical and human health impacts were recorded. These
included for example habitat medication, disruption of ecosystem functioning,
competition with or predation on native species, economic impacts arising from fouling
to disruptions to aquaculture, and human health impacts such as potential allergic
reactions and fatal incidents.
5.2.2 Findings
In total, 2,552 species with an invasion history were collated in the database. Based
on the above searches, introduction pathways were determined for 87% of the
species and about a fifth (approximately 400) could be ballast mediated (Figure 5.3).
250 were recorded from the tropics only or are considered cosmopolitan, and 91 of
these were recorded from the Sunda Shelf province.
Figure 5.3 Ballast-mediated aquatic species with an invasion history recorded from the six
sources
Blacklist
Forty ballast-mediated species reported to be present in Singapore are included in
the blacklist. Of these, 13 are native, 10 are alien and 17 cryptogenic and include:
two dinoflagellate Gymnodinium catenatum and Gymnodinium
impudicum;
two molluscs Mytilopsis sallei and Brachidontes striatulus;
the polychaete Hydroides sanctaecrucis
MEPC 70/INF.21
Annex, page 55
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
the goby Yongeichthys virgatulus
Only G. catenatum and M. sallei are recorded by (Nparks, 2012_Unpublished) as
being invasive based on their invasion history and being listed in publicly accessible
invasive species databases. The other four species may be regarded as introduced
species as they are not listed in the used database and there is little supporting
literature indicating substantial economic impact.
Watch List and Priority List
Depending on whether the species were considered for the Johor Strait or for the
Singapore Strait area where salinity ranges and temperature may be slightly different,
the report defined 22-24 species as high risk, 29-30 species as mid-risk, and 18-19
species as low risk. The high-risk species have been consolidated as priority-listed
species. Eventually, the priority listed species with the greatest potential impacts were
assessed to be the dinoflagellate Alexandrium minutum, the jellyfish Aurelia aurita,
and the tubeworm Ficopoomatus engimaicus.
5.2.3 Conclusions
The present study from Nparks, together with the comprehensive marine biodiversity
surveys taking place in Singapore contribute substantially towards setting an initial
baseline and a starting point for the identification of target species to be taken in
consideration under the SRA approach to exemption. In particular, the flowchart of
the data mining process outlines the steps involved in creating a blacklist and watch
list to help focus modelling exercises on high-risk species.
Whilst the study was only conducted at a local scale for Singapore alone, the
methodology could easily be extended to other countries in the proposed SRA and
scaled up to the level of the EAS creating a multi-scale analysis.
Although it is recommended that individual Parties to the SRA carry out their own
identification of target species, the database developed by Nparks
(2012_Unpublished) already provides a Black List and Watch List for the Sunda Shelf,
i.e. from the first three steps in Figure 5.2. This could be used as the starting point for
an SRA assessment for the EAS region.
However, there are gaps in the biodiversity records for tropical Southeast Asian. This
may potentially cause difficulties in generating a robust list of target species for the
Risk Assessment under the G7 Guidelines. Until the regional integration of data is
optimal, the present approach to the selection of target species can be considered as
appropriate in a SRA approach because the potential lack of biodiversity reports can
be partially compensated by proper hydrodynamic studies which give a good
estimation of the potential dispersion of non-reported species (simple particle tracking
models for instance).
5.3 Hydrodynamic modelling and Agent-Based Modelling
The examples provided in Section 4 that outlined the evidence base for connectivity
in the EAS Region provided some considerations for the detailed HD modelling and
ABM that is needed to define and confirm the boundaries of the SRA.
A number of tools exist for analysing the dispersal of planktonic or pelagic stages or
marine organisms, as follows:
MEPC 70/INF.21
Annex, page 56
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
The DHI suite of MIKE Software, for example, the 2D flow model MIKE
21 FM Hydrodynamic (HD) and Particle Tracking (PT) modules;
The ESA/Brockmann Consult DUE Innovator II BWE risk assessment
model which assesses the impacts of ballast water exchange and takes
drift into account (www.brockmann-consult.de);
A prototype quantitative biophysical model of the North Sea region to
model the risk of invasive species spread by ballast water developed by
DHI (Hansen et al., 2014);
The FETM-ERSEM model describing biological and physical-chemical
processes and their interactions in both the water column and on the
bottom, developed by a NIOZ-lead consortium (www.nioz.nl/ersem-
getm);
An interface for online larval dispersal based on archived currents from
oceanographic models and particle tracking techniques developed by
CSIRO (www.csiro.au/connie2/);
The use of a MARS 2D mathematical model developed at IFREMER to
evaluate dispersion of particles (Masson et al., 2013).
5.3.1 HD model
Where initial modelling indicates zones of low connectivity or eddies or other
hydrodynamic features, a multi-scale analysis of connectivity would be advisable. For
example, the Gulf of Thailand with low flushing rates and high residence times may
have a lower degree of natural ecological connectivity to other areas of the EAS
region. A multi-scale analysis can be achieved through the nested boundaries of the
IMO Globallast regions and subzones with some further nesting to capture finer local
scales as illustrated in the coral and sea start dispersion studies. This can be reflected
in the resolution of the mesh for the models at different scales.
Modelling can be carried out on a 2-dimensional basis where the surface, mixed layer
is vertically homogeneous, for example due to shallow water depths. However, under
some conditions, for example, on the edge of the continental shelf and in deep water
the stratification and wind effects are important and a 2D model may under predict
the surface currents. Examples in Section 4 illustrate the role that small scale (time
and space) hydrodynamic tidal patterns and geographical barriers such as islands
can play in determining larval dispersal and successful settlement. In such situations,
there will be a strong argument for the inclusion of higher resolution 3D modelling to
describe both deep ocean effects and nearshore tidal forces to fully describe the
patterns of dispersal.
It is further important to choose parameters carefully so that the patterns of interest
are not missed or disappear into seasonal averages.
5.3.2 ABM
The identification of the target species under Step Two of the SRA workflow provides
information on:
1. the scale over which the target species are likely to naturally
disperse and hence the scale for the modelling; and
MEPC 70/INF.21
Annex, page 57
https://edocs.imo.org/Final Documents/English/MEPC 70-INF.21 (E).docx
2. the timeframe over which the target species are likely to disperse in
a single event and hence the timeframes for the modelling.
As indicated by the case studies in Section 4, different parameters will be needed to
model the dispersal of pelagic larvae of benthic species (that might have an average
larval life span of only 90 days on average) compared to the dispersal of plankton
such as copepods that reproduce and move continually as long as they don't reach
any kind of barrier (including seasonal variation) that prevents them from growing
(e.g. moving out of their temperature range).
The scale over which the target species can disperse was discussed in Section 3.5
and estimated to range from 25 to 50.7 km per year according to the studies quoted.
This information could be used to help set the initial boundaries for the SRA study and
the time frames for the analysis. For example, if among the selected target species
the shortest distance that a pelagic species will disperse in a year is 25 km, then a
five year analysis may take in a spread of 125 km.
The time span allowed for the natural spreading of the species is a defining parameter
for the risk assessment in the context of an SRA. Too short a time span, and an
organism will not have sufficient time to travel (whether by pelagic larval drift or adult
swimming). Too long a time span, and this would result in an overly lenient
assessment. This was also reflected by comments from Canada in MEPC 69/4/15:
"Considering natural dispersion over very long time periods will tend to de-emphasize
the importance of shipping as a vector for increasing the spread of harmful aquatic
organisms and pathogens relative to natural forces. On the other hand, the use of
short periods will exaggerate the role of ships in spreading species that would likely
spread naturally in due course."
MEPC 69/INF.25 has suggested three options that may be included in the
considerations when defining the criterion of SRA delineation:
5 years. The duration of the G7 IMO exemption scheme is five years
with intermediate review;
10 years. The time for projecting the dispersal could be 10 reproductive
seasons or maximum 10 years; or
16 years. On a purely administrative basis, according to the Biological
Opinion of the US National Marine Fisheries Service, 16 years is allowed
to establish the effects of invaders (2 x 8 years).
Clearly, the i