APPENDIX III - Meeting Documents

Meeting Document 02/3 Background Information Global Environmental Facility - Funding - Marine Programmes

What is SGP?

The UNDP/GEF Small Grants Programme (UNDP/GEF-SGP)

GEF SGP “at a glance”

• Funded by the Global Environment Facility (GEF), implemented by UNDP and executed by UNOPS
• 63 countries participate in the Programme
• 53 country offices and a headquarters office staffed by 5 people
• Participating countries have ratified the Conventions on Biological Diversity and Climate Change
• Funding to date: US$117.35 million and US$65.66 million by other partners in cash or in-kind equivalent
• Maximum grant amount per project: US$50,000
• Nearly 3000 projects funded to date: 60% biodiversity, 15% climate change, 6% international waters, 19% multi-focal
• Benefits have reached more than 4000 communities worldwide

Established in 1992, the year of the Rio Earth Summit, The GEF Small Grants Programme [SGP] embodies the very essence of sustainable development. By providing financial and technical support to projects in developing countries that conserve and restore the natural world while enhancing wellbeing and livelihoods, SGP demonstrates that community action can maintain the fine balance between human needs and environmental imperatives. SGP links global, national and local-level issues through a transparent, participatory and country-driven approach to project planning, design and implementation. Grants are made directly to non-governmental organizations [NGOs] and community-based organizations [CBOs] in recognition of the key role they play as a resource and constituency for environment and development concerns. Though SGP grants are small, their impact is large. More than 3000 projects in 63 countries have addressed adverse environmental changes and enriched the lives of tens of thousands of people, in Africa, Near East, Asia and the Pacific, Europe and Latin America and the Caribbean.

The SGP recognizes that the destruction of ecosystems and the species that depend upon them; increasing levels of carbon dioxide and other greenhouse gases in our atmosphere; pollution of international waters: are the life-threatening challenges we face. Though environmental decay endangers all of us, poor people are most at risk because they depend on access to natural resources for their livelihoods and because they live in fragile ecosystems. The programme operates on the premise that people will be empowered to protect their environment when they are organized to take action, have a measure of control over access to the natural resource base, have the necessary information and knowledge, and believe that their social and economic wellbeing is dependent on sound long-term resource management.

However, SGP is more than simply a fund that provides small grants to improve the local environment. By raising public awareness, building partnerships, and promoting policy dialogue, the SGP seeks to help create a more supportive environment within countries for achieving sustainable development and addressing global environment issues. Six hundred organizations worldwide have provided co-funding and other forms of collaboration, including significant contributions by the UN Foundation, the European Commission and the Government of the Netherlands. SGP is mandated to raise project co-financing that matches GEF funds.

The principle objectives of the Small Grants Programme are to:

• Demonstrate community-level strategies and technologies that could reduce threats to the global environment if they are replicated over time;
• Draw lessons from community-level experience, and support the spread of successful community-level strategies and innovations among CBOs and NGOs, host governments, development aid agencies, the GEF, and others working on a larger scale;
• Build partnerships and networks of local stakeholders to support and strengthen community, CBO, and NGO capacity to address environmental problems and promote sustainable development.

The decentralized structure of the Small Grants Programme encourages maximum country and community level ownership and initiative.

SGP Participating Countries

Countries currently participating in the programme include:

AFRICA

• Botswana
• Burkina Faso
• Côte d'Ivoire
• Ghana
• Kenya
• Mali
• Mauritania
• Mauritius
• Namibia
• Senegal
• South Africa
• Tanzania, United Republic of
• Uganda
• Zimbabwe

ARAB STATES

• Egypt
• Jordan
• Lebanon
• Morocco
• Palestianian Authority
• Tunisia

ASIA AND PACIFIC

• Bhutan
• India
• Indonesia
• Iran, Islamic Republic of
• Malaysia
• Mongolia
• Nepal
• Pakistan
• Papua New Guinea
• Philippines
• Sri Lanka
• Thailand
• Viet Nam

EUROPE

• Albania
• Kazakhstan
• Krygystan
• Lithuania
• Poland
• Turkey

LATIN AMERICA AND CARIBBEAN

• Barbados (including Anguilla, Antigua, British Virgin Islands, Dominica, Grenada, Montserrat, St. Christopher and Nevis, St. Vincent and the Grenadines, St. Lucia)
• Belize
• Bolivia
• Brazil
• Chile
• Costa Rica
• Dominican Republic
• Ecuador
• El Salvador
• Guatemala
• Honduras
• Mexico
• Peru
• Trinidad and Tobago (including Suriname).

In each participating country, a broad-based national steering committee (NSC) provides overall guidance and strategic direction for the programme and screens and selects projects for grant awards. The committee guides the development of a country strategy and establishes country-specific eligibility criteria within the framework of the overall GEF guidelines. Members of the NSC serve on a voluntary basis and typically represent the government (which must endorse the programme); the CBO/NGO community; national academic, scientific, and technical institutions; and the UNDP Country Office.

A national coordinator is responsible for managing the implementation of the country programme. The coordinator works in close partnership with CBOs and NGOs to help them formulate their project proposals, visits the sites of proposed activities, supports the work of the NSC, and ensures sound programme monitoring and evaluation.

Eligible Activities

GEF/SGP grants are awarded for activities which support community-level action in the biodiversity, climate change, and international waters focal areas. Activities that address land degradation issues-primarily concerning desertification and deforestation-can be supported if they relate to one or more of these focal areas.

To be eligible for GEF/SGP support, a project proposed for funding must fit the GEF/SGP country programme strategy and country-specific eligibility criteria approved by the NSC. It must also be consistent with the Operational Strategy and relevant Operational Programs established by the GEF:

• In the biodiversity focal area, activities must promote the conservation and sustainable use of biological resources in arid and semi-arid ecosystems; coastal, marine and freshwater ecosystems; forest ecosystems; or mountain ecosystems.
• In the area of climate change, activities must either demonstrate the removal of local barriers to energy conservation and energy efficiency, or promote the adoption of renewable energy.
• In the international waters focal area, activities must address environmental concerns in a specific waterbody shared by two or more countries (such as freshwater drainage basin that is regionally significant or a large marine ecosystem), or address land-based threats to international waters.

Given the focus on small-scale, community-level initiatives, the UNDP/GEF/SGP is not active in the ozone layer focal area. Several different kinds of activities are eligible for funding under the UNDP/GEF Small Grants Programme:

• Community-based assessment and planning (planning grants): Small amounts of grant funds (typically no more than US$2,000) are available to support pre-project participatory assessment and planning activities designed to strengthen community participation in project identification and development.
• Pilot demonstration activities: Most funded projects are activities that test and demonstrate the viability of innovative community-level approaches to global environmental problems.
• Capacity development: Although most demonstration projects include capacity development components, grants may be awarded for targeted technical assistance and training activities which focus on developing CBO and NGO capacities in the GEF focal areas.
• Monitoring and analysis: Grants funds may also be made available to intermediary NGOs and research centers (including universities) to support programme monitoring; to help identify, assess, and document best practices; and to prepare case studies of UNDP/GEF/SGP-supported projects. The use of participatory methods in monitoring and analysis activities is encouraged.
• Dissemination, networking, and policy dialogue: In order to leverage UNDP/GEF/SGP project experience, grant funds are available to support dissemination of innovations and best practices, relevant networking activities, and policy dialogue efforts aimed at promoting a supportive policy environment for community-level action in the GEF focal areas.

Characteristics of UNDP/GEF/SGP-Supported Projects

In addition to meeting the basic GEF criteria, priority is given to activities that:

• Facilitate community participation in their design, implementation, and evaluation.
• Improve household and community access to productive assets and income/employment opportunities.
• Consider gender roles and relationships, with a focus on the needs of women.
• Recognize the roles and importance of indigenous knowledge and resource management systems, and of local institutions and patterns of social organization.
• Support approaches to technology adoption and diffusion that engage stakeholders directly and take into account their needs and priorities.
• Expand community access to local sources of technical assistance and training.
• Include local contributions and cost-sharing by stakeholders.

HOW TO APPLY FOR A GRANT

National and local NGOs and CBOs may propose projects for grant support under the Small Grants Programme.Procedures for project proposal screening and approval are generally as follows:

1. The project proponent contacts the SGP national coordinator to receive project application guidelines and forms.

2. With assistance from the national coordinator and using the standard SGP format, the proponent prepares a brief project concept paper and submits this to the coordinator.

3. The national coordinator reviews and pre-screens the concept paper according to GEF criteria and criteria adopted by the NSC for activities in that country.

4. If the project is judged eligible, the project proponent prepares a project proposal; in some cases, this step may be supported by a planning grant.

5. Completed project proposals are submitted by the national coordinator or the NSC.

6. The NSC reviews the proposal and either accepts it, rejects it, or returns it to the proposer with a request that further work be done on formulating and refining the project data.

7. Approved proposals enter the national UNDP/GEF-SGP work programme. UNDP/GEF-SGP grants are usually paid in three installments: an up-front payment to initiate the project; a midterm payment upon receipt of a satisfactory progress report; and a final payment on receipt of a satisfactory project completion and final report.

BIODIVERSITY PLANNING SUPPORT PROGRAMME

Introduction

The Biodiversity Planning Support Programme is a multi-donor initiative implemented by the United Nations Development Programme (UNDP) and the United Nations Environment Programme (UNEP) with core financing from the Global Environment Facility. The Governments of Norway and Switzerland also provide co-financing.

Purpose of the Programme

The programme was established to respond to needs recognized by the Parties to the United Nations Convention on Biological Diversity for strengthening national capacity to prepare and implement National Biodiversity Strategies and Action Plans (NBSAPs) in compliance with Article 6 of the Convention.

Programme Objectives

• To support the application of regional expertise toward assisting countries to complete high quality NBSAPs;
• To assist national planning teams to achieve effective integration of biodiversity conservation and sustainable use into sectoral planning and national policy; and
• To ensure timely and cost-effective implementation of biodiversity projects to address national priorities.

The programme has three components that will be implemented at global and regional level:

Information Gathering and Dissemination

Specialized information on biodiversity planning and issues related to the CBD will be compiled, translated as appropriate, and distributed to national planning teams. The programme will establish user-friendly mechanisms at regional level to foster regular and ongoing information exchange including web sites, electronic mail list-servers, and help lines. The global Website, to become operational in mid 1999, can be found at: http://www.undp.org/gef.

Guidelines and Best Practice Experience

Under the leadership of UNEP, the programme will develop guidelines, training modules and facilitate dissemination of “best practice” experience developed during the course of NBSAP preparation. The programme will give priority attention to issues emerging from National Reports and following guidance from the Conference of Parties to the CBD. Thematic issues to be addressed by the programme will include:

• Mainstreaming biodiversity into sectoral and economic policy development and planning;
• Developing financial strategies for NBSAP implementation;
• Developing incentive measure for biodiversity conservation and sustainable use;
• Facilitating a harmonized and integrated approach for the implementation of the objectives of the CBD with other biodiversity-related conventions.

Regional Exchange and Thematic Workshops

The Biodiversity Planning Support Programme will organize Regional Exchange and Thematic Workshops to promote intra-regional and global exchange of knowledge, experience and expertise. UNEP will take lead responsibility in organizing Thematic Workshops and will also coordinate information exchange workshops in Africa. UNDP will have primary responsibility for organizing and coordinating workshops for exchange of information and experience in the other regions. The programme will include the following:

• Africa: Regional workshop and Thematic workshops
• Arab States: Regional workshop and Thematic workshops
• Asia: Regional workshop and Thematic workshops
• Pacific Island Countries: Regional workshops
• Eastern Europe and the CIS: Regional workshop and Thematic workshops
• Caribbean: Regional workshops
• Latin America: Regional workshop and Thematic workshops
• Small Island States: Thematic workshop

As thematic issues emerge from the COP guidance and national practice, global workshops will be organized to provide synthesis of analytical expertise and practical experience. If your country is interested in hosting or participating in these workshops, please contact the Biodiversity Planning Support Unit.

Programme Oversight

The Project Advisory Committee will oversee the programme activities. The committee is co-chaired by UNDP and UNEP and includes representatives from the GEF and CBD Secretariats, donor countries, regional experts, and international NGOs. To ensure timely feedback from crucial stakeholders, an Advisory Panel will be established as an open forum for two-way dialog on programme implementation and performance.

For more information, please contact:
Dr. G. Ken Creighton
Biodiversity Planning Support Programme
UNDP-GEF 304 East 45^th Street, 1606
New York, New York 10017 USA
Tel: 1-212-906-6757
Fax: 1-212-906-6568
Email: [email protected]
Ms. Carmen Tavera
Biodiversity Enabling Activities
UNEP, Nairobi, Kenya
Tel: 254-2-624182
Fax: 254-2-624268
Email: [email protected]

ELIGIBILITY CRITERIA & PROJECT CYCLE

Any eligible individual or group may propose a project, which must meet two key criteria: It must reflect national or regional priorities and have the support of the country or countries involved, and it must improve the global environment or advance the prospect of reducing risks to it. GEF project ideas may be proposed directly to UNDP, UNEP, or the World Bank.

Country eligibility to receive funding is determined in two ways. Developing countries that have ratified the relevant treaty are eligible to propose biodiversity and climate change projects. Other countries, primarily those with economies in transition are eligible if the country is a party to the appropriate treaty and is eligible to borrow from the World Bank or receive technical assistance grants from UNDP.

Meeting Document 02/4 - South African Commercial Catch Data

FISHERIES DATA PROVIDED BY SOUTH AFRICA (y = fish discharge, N = no discharge) Declared fish (product weight in kg) aboard foreign vessels visiting South African ports

Meeting Document 02/5 A Summary of Japanese Trawl Fishery Newly Developed in the Southwest Indian Ocean (FAO Area 51)

Mitsuo SAKAI and Shigeyuki KAWAHARA
National Research Institute of Far Seas Fisheries, Fisheries Agency
5-7-1, Orido, Shimizu, Shizuoka, 424-8633 Japan
<[email protected]>

For the past 20 years, few Japanese distant-water trawlers have operated in the Southwest Indian Ocean (FAO Statistical Area 51) until a recent deepwater trawl fishery newly developed in the open sea. Two stern trawlers of about 1150GRT, have started fishing operations targeting mainly alfonsino (Beryx aplendens) and orange roughy (Hoplostethus atlanticus) since March 2001. The cod-end of 60 to 120 mm mesh size was used. The fishing grounds were developed around the seamounts located in the area 35 to 45°S and 45 to 55°E along the Southwestern Indian Ridge at the depth range from 400 to 1700m.

One of the trawlers started operations from 18 March and the other started from 2 April 2001. Nominal catches, fishing efforts and catch per effort of the newly started Japanese trawl fishery in 2001 are shown in Table 1. Number of days fished and tows, and total time towed (hrs) were 314 days, 464 tows, and 1435 hours, respectively. Fishing efforts increased in April and a higher level of effort was maintained until July (Figure 1). Total catch was 4143.9t, and alfonsino and orange roughy account for major part of the catch, 2904.3t (70.1%) and 410.9t (9.9%), respectively. Catches of smooth oreo dory (Pseudocyttus maculates) were next in harvest abundance. Highest monthly catches for alfonsino were recorded in April and thereafter the alfonsino catch decreased gradually (Figure 2). Catch per unit effort (per hours and per tows) for alfonsino is shown in Figure 3. The CPUE seems to be slightly declining since the start of the fishing in March but the trend is not clear. Orange roughy were caught principally from April to July. Some catches of armourhead or boarfish (Pentacerotidae; Pseudopentaceros sp.), black dory (Allocytus niger) and cardinalfish (Epigponus sp.) were mainly observed from March to July.

Figure 1 Monthly fluctuation of nominal fishing effort (duration of tow, hrs) of Japanese deepwater trawler in 2002

Figure 2 Monthly changes of nominal catches of Alfonsino (Beryx aplendens)

Figure 3 Monthly nominal CPUE fluctuation for Alfonsino (Beryx aplendens)

The species and species groups mentioned above were tentatively determined because samples are not available for species identification until now. At present our research efforts are limited to collecting catch-data statistics and information on the fishery. Further participation by the Japanese fishery in the study area would facilitate collecting more appropriate biological data of the concerned species as the next step to enable the stock assessment in the future.

Table 1
Nominal catches and fishing efforts of Japanese trawl fishery in FAO Area 51 in 2001

1) Specie and species groups, especially their scientific names, are preliminary ones.

2) Greeneye (Chlorophthalmidae), bluenose (Hyperoglyphe antarctica), barracouta (Thyrsites atun), sharks, shrimps/lobsters, and squids.

Meeting Document 02/6 - A Summary of the 1977 and 1978 Cruises by the Jamarc Trawl Survey on the Deepwater Fisheries Resources of the Southwest Indian Ocean

Mitsuo SAKAI
National Research Institute of Far Seas Fisheries, Fisheries Agency
5-7-1, Orido, Shimizu, Shizuoka, 424-8633
<[email protected]>

Scientific research surveys on the development of deepwater fisheries resources were made by the JAMARC (Japan Marine Fishery Resources Research Center) in the Southwest Indian Ocean from April 1977 to March 1979 with two stern trawlers, the Shinkai Mam and the Dai 2 Ryuyou Mam, which were of 3000GRT (JAMARC 1978a,b; JAMARC 1980a,b). The research areas (10-50°S and 40-70°E) were mainly on banks, shelves and slopes of islands: Saya de Malha Bank, Mdagascar Bank, East Marion Bank, Ob and Lena Bank, Kerguelen Island, Crozet Island. The surveys examined their feasibility as demersal fishing grounds, collected topographic information data on the distribution of main species, and also the possible development of cod icefish (Nototheniidae) and mackerel icefish (Channichthyidae) as fisheries resources.

In Saya de Malha Bank, mackerel (Carangidae), lizardfish (Synodontidae), and butterfly breams (Nemipteridae; Nemipterus sp.) were the dominant species in the range of towed depths, 40 to 300m. On the Madagascar Bank, snappers (Lutjanidae; Etelis sp.) and jacks were distributed under 200m in depth, and boarfish (Histioptenidae; Pseudopentaceros sp.) and stargazers (Uranoscopidae; Uranoscopus sp.) were found in waters over 700m in depth. In Crozet Islands, and Ob and Lena Banks, toothfish (Nototenidae; Dissostichus spp.) and cod icefish (Nototenidae; Notothenia spp.) dominated. In Kerguelen Islands, Antarctic cod (Nototenidae; Notothenia rosii) dominated over 50% of the catch, and mackerel icefish (Channichthyidae; Champsocephalus gunnari) and unicorn icefish (Channichthyidae; Channichthys rhinoceratus) were the next most abundant species.

Two survey areas, Saya de Malha Bank (10-12 °S and 60-62 °E) and Mdagascar Bank (33 °S and 43 °E), are classified in the relevant FAO Statistical Area 51. Survey results, including catches and effort in the survey areas, are summarized in Table 1. Few catches of the relevant species [orange roughy (Hopolostethus atlanticus) and alfonsino (Beryx splendens)] were observed in these areas.

LITERATURE CITED

JAMARC 1978a. Cruise report 1977 on the new deepwater fisheries resources for developing as commercialized ones in the Southwest Indian Ocean. Report 1977, No. 12 (in Japanese).

JAMARC 1978b. Cruise report 1977 on the new deepwater fisheries resources for developing as commercialized ones in the Southwest Indian Ocean (Data). Report 1977, No. 12-1 (in

Japanese). JAMARC 1980a. Cruise report 1978 on the new demersal fisheries resources for developing as commercialized ones in the Southwest Indian Ocean. Report 1978, No.7 (in Japanese).

JAMARC 1980b. Cruise report 1978 on the new demersal fisheries resources for developing as commercialized ones in the Southwest Indian Ocean (Data). Report 1978, No.7-1 (in Japanese).

Table 1

Resume of catch and effort on the past scientific cruises conducted by Japanese Government (JAMARC) in FAO Statistical Area 51 of the Southwest Indian Ocean

Data Forms
Country Catches and Vessel Catches

Country: Japan Year: 2001

Meeting Document 02/7 - Past USSR - Ukrainian Fishing Survey and Exploratory Activities

Evgeny Romanov
World Ocean Fisheries Resources Dept.
Southern Scientific Research Institute of Marine Fisheries and Oceanography (YugNIRO)
2, Sverdlov Str
Kerch 98300
Crimea, Ukraine
<[email protected]>

Fisheries and research surveys of seamounts in the Southern Indian Ocean were carried out using vessels of 6 types: PPR, RTMS, BMRT, RTMA, SRTM, SRTMK. Their principal specifications of the vessels are presented in Table 1.

Table 1
Principal characteristics of Ukrainian research and fishing vessels operating in the Southern Indian Ocean

Research vessels have the same fishing abilities as commercial fishing vessels, except less hold capacity as a part of the vessel is allotted to scientific laboratories. Vessels of the SRTM and SRTMK type were generally used for line, longline and pot fishing. For this purpose they were specially fitted for line fishing, lobster pot fishing, bottom and pelagic long-line fishing, or lift nets fishing. However their trawl fishing ability was maintained.

Statistics of fishing effort and catches of Soviet (Ukrainian) fleet by vessel types and fishing gears are presented in Tables 2 and 3. Target species for the Soviet (Ukrainian) fleet were alfonsino, rubyfishes (two species), and butterfishes (two principal species). The Soviet (Ukrainian) fishing fleet never targeted orange roughy and dories in this area although their capture during research cruises was recorded.

Table 2 Catch statistics of Soviet/Ukrainian vessels on Southwestern Indian Ridge Seamounts (FAO Statistical Areas 51.01 and 51.02)

¹ For vessels of SRTM-type, data on fishing effort are unavailable (fishing effort consisted of hand lines, mechanized lines, pot fisheries, bottom longlines and sometimes trawl fisheries).

² Since exploratory fishing by bottom longline were carried out by chartered vessel from Russian home ports, their catches are not included in the catch statistic of the Ukraine.

Table 3 Catch statistics of Soviet/Ukrainian vessels at the Mid-Indian Ridge and Broken Ridge Seamounts (FAO Stat Area 51.04)

Meeting Document 02/8 - New Zealand High Seas Fishing Permit HS02-03/XXX/AAA

Pursuant to Section 113H of the Fisheries Act 1996:

NAME OF PERMIT HOLDER

Client number:

is hereby authorised to use the following vessel to take and transport fish, aquatic life and seaweed on the high seas, subject to the conditions of this permit, the Fisheries Act 1996, any relevant regulations made under that Act, and all other relevant legislation:

Period for which Fishing is Authorized

This permit is valid from 01 May 2002 until 30 April 2003 (inclusive).

Amendment and Revocation of Conditions

The Chief Executive may from time to time, by written notice to the permit holder, amend, add to, or revoke any of the conditions in this permit with effect from the date specified in the notice.

Chief Executive approvals and exemptions

Where any event or thing requires the approval or exemption of the Chief Executive, that approval or exemption may be given subject to conditions, that must be complied with as though they were conditions of this permit.

If required to obtain an approval or exemption from the Chief Executive under this permit, the permit holder must apply, in writing, to FishServe, PO Box 297, Wellington.

Definitions

Terms used in this permit have the meanings set out in Section 2(1) the Fisheries Act 1996 or as set out below:

“FCC” means the Fisheries Communication Centre of the Ministry of Fisheries.

“High seas returns” means high seas trawl catch effort returns, high seas squid jigging catch effort returns, high seas tuna longlining catch effort returns, and high seas catch effort landing returns.

“High seas trawl catch effort returns”, “High seas squid jigging catch effort returns”, “High seas tuna longlining catch effort returns”, “High seas catch effort landing returns” mean the forms that have been approved by the Chief Executive for that purpose.

“Landing” means the removal or discharge of fish, aquatic life, or seaweed from the vessel in respect of which this permit has been issued. A landing is also deemed to occur from the vessel when the vessel ceases to be registered or is re-registered under the Fisheries Act, for whatever reason and by whatever mechanism.

“Permit holder” means the holder of this high seas fishing permit issued under Section 113H of the Fisheries Act 1996.

“Restricted areas” means areas defined in conditions 45 and 46.

“Trip” means the movement of the vessel from a port and its return to that port or arrival at another port, whether the port is in New Zealand or elsewhere, for the purpose of taking or transporting fish, aquatic life, or seaweed pursuant to this high seas fishing permit.

“Written notice to FCC” means the transmission of the required information by means of either electronic mail (E-mail), or facsimile to FCC at the following address/number:

E-mail [email protected]

Facsimile +64 4 801 5381

GENERAL CONDITIONS

1 This permit or a copy of it must be carried on board the vessel when on a trip.

COMMUNICATIONS

2 The permit holder must ensure that a copy of all communications made to and received from FCC or any other New Zealand Government agency in relation to a trip, is kept on board the vessel during that trip.

3 The permit holder must ensure that the master of the vessel has a sound command of the English language.

Intention to leave port

4 When leaving on a trip, the permit holder must ensure that written notice to FCC of the following matters is provided at least 6 working days prior to the departure of the vessel from a port:

4.1 name of high seas permit holder,
4.2 vessel name,
4.3 international radio call sign,
4.4 name of vessel master,
4.5 intended date and time of departure from port (specify whether using UTC or NZST),
4.6 port of departure (including state and country, if outside NZ),
4.7 species, weight and state of bait (if any) on board vessel at time of departure from port,
4.8 area and species intended to be fished on trip,
4.9 intended method of fishing on trip,
4.10 proposed date of arrival at any port, and
4.11 name of port of arrival.

5 The permit holder must ensure that written notice to FCC is provided immediately if any details in condition 4 change at any time prior to or during a trip.

Notification of entry to/exit from New Zealand fisheries waters, any foreign fishing jurisdiction, or any restricted area

6 When on a trip, the permit holder must ensure that written notice to FCC of the following matters is provided immediately on the vessel’s entry to or exit from New Zealand fisheries waters, any foreign fishing jurisdiction, or any restricted area, including when transiting:

6.1 name of permit holder,
6.2 vessel name,
6.3 international radio call sign,
6.4 date, time and position of crossing of boundary (specify whether using UTC or NZST), and
6.5 estimated time, date and position of exit from New Zealand fisheries waters, if entering New Zealand fisheries waters to transit.

Notification of intention of entry to port

7 When on a trip, the permit holder must ensure that written notice to FCC of the following matters is provided no later than 48 hours prior to the arrival of a vessel in any port (New Zealand or elsewhere):

7.1 name of permit holder,
7.2 vessel name,
7.3 international radio call sign,
7.4 estimated date and time of arrival in port (specify whether using UTC or NZST),
7.5 intended port of call,
7.6 an estimate of the species, weight and state of fish, aquatic life, or seaweed on board the vessel,
7.7 estimated date and time of commencement of unloading (if any),
7.8 species, weight and state of fish, aquatic life, or seaweed to be landed (if any), and
7.9 if landing to a port outside New Zealand fisheries waters, date of issue of approval granted under condition 19 of this permit, or a request to obtain approval.

OBSERVERS

8 The permit holder is required to carry an observer on a trip if requested to do so by the Chief Executive.

9 The permit holder must meet all costs arising from carriage of an observer if requested to do so by the Chief Executive.

10 The provisions relating to the placement of observers are contained in Part XII of the Fisheries Act 1996.

INSPECTION

11 The permit holder must ensure that a fishery officer or observer inspects the vessel and certifies that the vessel is empty of fish, aquatic life and seaweed prior to the vessel departing from a New Zealand port on a trip, unless prior written exemption has been obtained from the Chief Executive.

12 Condition 11 does not prevent the carriage of bait to be used for high seas fishing, provided the species, weight and state of the bait is recorded and notified in accordance with condition 4 prior to port departure.

VESSEL MONITORING

13 The permit holder must ensure that the vessel carries and operates an automatic location communicator currently registered with the Ministry of Fisheries at all times during the term of this permit.

14 The permit holder must comply with the requirements of the Fisheries (Satellite Vessel Monitoring) Regulations 1993 and any circulars issued thereunder as if they were conditions of this permit. References in those regulations to “the operator and master of any vessel required by these regulations to carry and operate an ALC” must be read as references to “high seas fishing permit issued under Section 113H of the Fisheries Act 1996”.

15 The permit holder must meet all costs arising from ensuring compliance with the requirements of the Fisheries (Satellite Vessel Monitoring) Regulations 1993.

VESSEL MARKINGS

16 The permit holder must ensure that the vessel is marked in accordance with the Fisheries (Commercial Fishing) Regulations 2001.

17 The permit holder must ensure that all tenders are clearly and legibly marked on at least one side of the hull with the international radio call sign of the vessel to which it is a tender.

LANDING AND OTHER DISPOSAL OF FISH

18 Each landing to a Licensed Fish Receiver and each transhipment within New Zealand fisheries waters must be supervised by a fishery officer or observer, unless otherwise advised by FCC. All associated costs of the supervision are to be met by the permit holder.

19 No fish, aquatic life, or seaweed may be landed to a port outside New Zealand fisheries waters without the prior written approval of the Chief Executive. See Schedule 1 for further information.

20 No fish, aquatic life, or seaweed may be transhipped while in a port or on a trip, either to, or from the vessel, whether on the high seas or otherwise, without the prior written approval of the Chief Executive. See Schedule 1 for further information.

REPORTING

21 For the purposes of the reporting conditions in this permit, the term “trip” does not include a trip that is solely for the purposes of transporting fish, aquatic life or seaweed pursuant to this permit.

22 Every permit holder who is required to complete high seas returns and catch landing returns must complete such returns in accordance with the explanatory notes attached to the returns and the requirements of the Fisheries (Reporting) Regulations 2001 as if they were conditions of this permit.

High seas returns

23 Every permit holder who takes fish, aquatic life, or seaweed by the method of trawling must complete high seas trawl catch effort returns.

24 Every permit holder who takes squid by way of jigging must complete high seas squid jigging catch effort returns.

25 Every permit holder who targets tuna by the method of longlining must complete high seas tuna longlining catch effort returns.

26 Every permit holder who is required to complete high seas trawl catch effort returns, high seas squid jigging catch effort returns, or high seas tuna longlining catch effort returns must-

26.1 complete such returns for each day or part day that the vessel is on a trip (including days where no fish, aquatic life, or seaweed is taken); and
26.2 furnish such returns to FishServe no later than 7 days after the last day of the trip.

Catch landing returns

27 Every permit holder -

27.1 who is required to complete high seas trawl catch effort returns, high seas squid jigging catch effort returns, or high seas tuna longlining catch effort returns; and

27.2 who lands any fish, aquatic life, or seaweed to a licensed fish receiver in New Zealand - must complete catch landing returns in respect of all landings for that trip (whether such landings occurred within New Zealand fisheries waters or elsewhere).

28 Every permit holder required to complete catch landing returns must do so in the following manner:

28.1 In respect of fish, aquatic life, or seaweed landed to a licensed fish receiver, the permit holder must complete catch landing returns immediately on landing, with the exception of the last 2 columns of the section of the return headed “Catch Landing Data” which must be completed immediately upon receipt of the necessary information required from a licensed fish receiver; and

28.2 In respect of fish, aquatic life, or seaweed landed other than to a licensed fish receiver, the permit holder must complete catch landing returns immediately on landing.

29 Every permit holder required to complete catch landing returns must furnish such returns to FishServe no later than 7 days after the last day of the trip.

High seas catch effort landing returns

30 Every permit holder who takes fish, aquatic life, or seaweed pursuant to this permit but who is not required to complete high seas trawl catch effort returns, high seas squid jigging catch effort returns, or high seas tuna longlining catch effort returns must complete high seas catch effort landing returns.

31 Every permit holder who is required to complete high seas catch effort landing returns must complete the section of the return headed “Catch/Effort Data” for each day or part day that the vessel takes fish, aquatic life, or seaweed.

32 Every permit holder -

32.1 who is required to complete high seas catch effort landing returns; and

32.2 who lands any fish, aquatic life, or seaweed to a licensed fish receiver in New Zealand - must complete the section of the return headed “Catch Landing Data” in respect of all landings for that trip (whether such landings occurred within New Zealand fisheries waters or elsewhere).^[9]

33 Every permit holder who is required to complete the section headed “Catch Landing Data” of the high seas catch effort landing return must do so in the following manner:

33.1 In respect of fish, aquatic life, or seaweed landed to a licensed fish receiver, the permit holder must complete the section of the return headed “Catch Landing Data” immediately on landing, with the exception of the last 2 columns which must be completed immediately upon receipt of the necessary information required from a licensed fish receiver; and 33.2 In respect of fish, aquatic life, or seaweed landed other than to a licensed fish receiver, the permit holder must complete the section of the return headed “Catch Landing Data” immediately on landing.

34 Every permit holder required to complete high seas catch effort landing returns must furnish such returns to FishServe no later than 7 days after the last day of the trip.

Other requirements

35 High seas return books can be obtained from FishServe.

36 An image of each type of high seas return is attached to this permit.

37 In addition, when taking fish, aquatic life, or seaweed pursuant to this permit, all the requirements of the Fisheries (Reporting) Regulations 2001 apply as if they were conditions of this permit, except:

37.1 High seas returns must be completed and furnished instead of:

37.1.1 catch, effort and landing returns;
37.1.2 trawl catch, effort and processing returns;
37.1.3 squid jigging catch, effort returns; or
37.1.4 tuna longlining catch, effort returns.

37.2 All times must be recorded in hours and minutes according to a 24 hour clock in UTC (Co-ordinated Universal Time) rather than New Zealand standard time or New Zealand daylight time.

37.3 The appropriate species code to be entered on a high seas return in respect of the particular species taken is specified on the Ministry of Fisheries Web site: www.fish.govt.nz

37.4 All fishstock codes must be reported as the appropriate species code followed by the area code “ET” (e.g. for orange roughy enter “ORHET”) unless a different area code is specified. An area-specific or species-specific authorisation or approval may require you to use a different area code when taking fish, aquatic life, or seaweed pursuant to that authorisation or approval. For example authorisations under the Fisheries (South Tasman Rise Orange Roughy Fishery) Regulations 2000 require holders to use the area code “STR” after the species code.

37.5 Where a species of fish, aquatic life, or seaweed is taken for which there is no corresponding species code, the species code for an unidentified species must be used (“UNX”). The permit holder must then obtain the correct scientific name for that species and report that scientific name to the Research Data Manager, Ministry of Fisheries (PO Box 862, Wellington) to enable the Ministry to create a code for future reporting of that species.

37.6 The latitude and longitude of each place where the species of fish, aquatic life, or seaweed were taken or where fishing commenced must be entered on the appropriate high seas return.

TRANSIT LIMITATIONS

Limitations within New Zealand fisheries waters

38 The permit holder must not take any fish, aquatic life, or seaweed within New Zealand fisheries waters during a trip unless written approval has been obtained from the Chief Executive prior to departing on that trip.

39 Where the vessel has departed from a New Zealand port on a trip, the vessel must proceed directly to the high seas unless an approval as specified in condition 38 has been obtained from the Chief Executive.

40 Where the vessel has entered New Zealand fisheries waters from the high seas, the vessel must proceed directly to port unless:

40.1 an approval as specified in condition 38 has been obtained from the Chief Executive; or
40.2 the vessel is transiting New Zealand fisheries waters.

Limitations within foreign fishing jurisdictions

41 The permit holder must not take any fish, aquatic life, or seaweed within any foreign fishing jurisdiction during a trip unless:

41.1 an approval to take fish, aquatic life, or seaweed within that foreign fishing jurisdiction has been obtained in respect of the vessel; and
41.2 prior to the trip, the permit holder has supplied a copy of that approval to the Manager, International and Biosecurity, Ministry of Fisheries, PO Box 1020, Wellington.

42 Where the vessel has departed from a port of any country other than New Zealand on a trip, the vessel must proceed directly to the high seas unless an approval as specified in condition 41 has been obtained and a copy of that approval has been supplied to the Ministry of Fisheries.

43 Where the vessel has entered any foreign fishing jurisdiction from the high seas or from another foreign fishing jurisdiction while on a trip, the vessel must proceed directly to port unless:

43.1 an approval as specified in condition 41 has been obtained and a copy of that approval has been supplied to the Manager, International and Biosecurity, Ministry of Fisheries, PO Box 1020, Wellington.; or
43.2 the vessel is transiting the foreign fishing jurisdiction.

GEAR RESTRICTIONS

Stowage of gear

44 Whenever the vessel is in an area where fishing, or certain types of fishing, are not permitted, the permit holder must stow the relevant fishing equipment in such a manner that it is not readily available for use for fishing.

AREA/SPECIES RESTRICTIONS

WestPac Bank

45 The vessel must not be used to trawl in the area enclosed by a line commencing at a point on the boundary of the New Zealand exclusive economic zone (EEZ) at 39⁰ S and 168⁰ 34’ E and proceeding due west to a point 39⁰ S and 166⁰ 30’E then proceeding due south to a point 40⁰ 30’ S and 166⁰ 30’ E then proceeding due east to a point 40⁰ 30’S and 167⁰ 24’E and then proceeding in a north-easterly direction to the point of commencement, unless the prior written approval of the Chief Executive has been obtained.

46 Unless the permit holder has the prior approval of the Chief Executive, the permit holder is not authorised under this permit to engage in fishing operations for fish stocks that are subject to measures established by the following organisations or arrangements:

• General Fisheries Commission for the Mediterranean (GFCM)
• Inter-American Tropical Tuna Commission (IATTC)
• International Baltic Sea Fishery Commission (IBSFC)
• International Commission for the Conservation of Atlantic Tunas (ICCAT)
• Indian Ocean Tuna Commission (IOTC)
• International Pacific Halibut Commission (IPHC)
• Northwest Atlantic Fisheries Organisation (NAFO)
• Northeast Atlantic Fisheries Commission (NEAFC)

47 The permit holder must not engage in fishing operations for anadromous fish stocks on the high seas.

Date:

Wayne Lowther
Registry Services Manager
Acting under delegated authority of the Chief Executive of the Ministry of Fisheries
Government of New Zealand

SCHEDULE 1

APPROVALS TO LAND OUTSIDE NEW ZEALAND FISHERIES WATERS AND TO TRANSHIP

Landings to ports outside New Zealand fisheries waters

Permit holders may obtain an approval to land to ports outside New Zealand fisheries waters by applying to the Chief Executive. Each approval may cover a number of foreign ports. When permit holders make their application for approvals, they therefore may wish to specify all the foreign ports that they may wish to land at. Approval applications should be sent to FishServe, PO Box 297, Wellington.

Transhipments

Permit holders may obtain an approval to tranship catch to, or receive catch from, other vessels while in a port or on a trip by applying to the Chief Executive. Each approval may cover a number of other vessels. When permit holders make their application for approvals, they therefore may wish to specify all the vessels that they may wish to tranship catch to, or receive catch from. Approval applications should be sent to FishServe, PO Box 297, Wellington.

SCHEDULE 2

STOCKS SUBJECT TO ARRANGEMENTS THAT NEW ZEALAND IS NOT A PARTY TO

Meeting Document 02/9^[10] - Predictive modelling of demersal fish distribution in the southern Indian and Southern Oceans

Bureau of Rural Sciences

AUSTRALIA

Final Report to the

Fisheries and Aquaculture Branch

September 2001

David Barratt and Richard Tilzey

© Commonwealth of Australia 2000

This work is copyright. The Copyright Act 1968 permits fair dealing for study, research, news reporting, criticism or review. Selected passages, tables or diagrams may be reproduced for such purposes provided acknowledgment of the source is included. Major extracts or the entire document may not be reproduced by any process without the written permission of the Executive Director, Bureau of Rural Sciences, PO Box 858, Canberra, ACT 2601.

The Bureau of Rural Sciences is an independent scientific agency within the Department of Agriculture, Fisheries and Forestry - Australia (AFFA).

Postal address:

Bureau of Rural Sciences
PO Box 858
Canberra ACT 2601

Internet: www.brs.gov.au

This publication does not represent professional advice given by the Commonwealth or any person acting for the Commonwealth for any particular purpose. It should not be relied on as the basis for any decision to take action or not take action on any matter that it covers. Readers should make their own further enquiries, and obtain professional advice where appropriate, before making any such decision.

The Commonwealth and all persons acting for the Commonwealth in preparing this publication disclaim all responsibility and liability to any person arising directly or indirectly from any person taking or not taking action based upon the information in this publication.

Acknowledgments

This project was largely funded from the Fisheries Resources Research Fund. We would like to sincerely thank all staff from Commonwealth and State fisheries authorities and agencies who have supplied data, advice and information, particularly John Garvey and Bob Stanley from the Australian Fisheries Management Authority, and Malcolm Clark from the New Zealand National Institute of Water and Atmosphere (South Tasman Rise data). The Australian high seas trawl fleet, in particular the Kailis and France Group, are thanked for permission to use their catch data records. This report benefited from comments made by Martin Exel, John Kalish and Kevin McLoughlin.

Executive Summary

Introduction

Since UNCLOS legislation pertaining to Straddling and Migratory Stocks was developed, Australia’s practical involvement with the management of finfish fisheries in international waters has largely been limited to pelagic fish species such as southern bluefin tuna (Thunnus maccoyii). For major pelagic species well-developed policies and multi-lateral international arrangements, such as the Convention for the Conservation of Southern Bluefin Tuna (CCSBT) and the Indian Ocean Tuna Commission (IOTC), are in place. However, from the late 1980s onwards, there has been increasing movement offshore to high seas waters outside Exclusive Economic Zones (EEZs) by demersal trawl fleets searching for deepwater species such as orange roughy (Hoplostethus atlanticus) oreos (Oreosomatidae) and alfonsino (Beryx splendens) that are often associated with seamounts and ocean ridges. Australian trawlers have been part of this movement. Orange roughy fisheries on seamount/ocean ridge features are often comparatively limited in area. In the Tasman Sea and nearby waters, Australian vessels have had to compete with New Zealand trawlers for such resources. Following the fish-down of orange roughy stocks within EEZs, the movement of trawlers offshore has increased in recent years. Thus, there is an increasing need to develop and implement sustainable management policies for demersal fisheries in international waters. In the absence of management controls, such fisheries are usually swiftly depleted.

In the Tasman Sea, high-seas orange roughy fisheries commenced on the North West Challenger and Challenger Plateaus and the Lord Howe Rise (Figure 1) in 1988. Another fishery on the Louisville Ridge, approximately 600 km east of the New Zealand EEZ, commenced in 1993. The South Tasman Rise fishery, which straddles the Australian EEZ (Figure 1), developed in 1997. The most recently discovered fishery (for orange roughy and other species, including oreos, alfonsino and armorhead or boar fish (Pseudopentaceros richardsoni) on the Madagascar Ridge in the Indian Ocean (Figure 3) was found by Australian and New Zealand vessels in late 1999. Vessels from several nations soon began to target this fishery and vessel numbers rapidly expanded from 7 in 1999 to >40 in 2000. Anecdotal reports estimate this fishery to have yielded around 50,000 t of fish by early 2001, but more recent catches have been low.

With the exception of the south Challenger Plateau and South Tasman Rise orange roughy fisheries (Figure 1) that are managed as straddling stocks, catches from high-seas demersal fisheries have been unregulated. Consequently, such fisheries typically go from ‘boom’ to ‘bust’ in comparatively short time, especially if the target species is a long-lived, slow-growing species, as is the case with orange roughy. For example, annual orange roughy landings from the Louisville Ridge fishery peaked at over 13,000 t in 1994/95, then swiftly declined to 1234 t in 1997/98. If high-seas demersal fisheries are to be fished in a sustainable manner, the development of management policies should be accorded a high priority. However, there is currently no legislative or policy framework (under the UNCLOS or other authority) by which such management can be implemented. A recent (November 1999) informal meeting between Australian, New Zealand, South African and other national officials discussed the possibility of developing a regional agreement (together with other interested nations) for southern Indian Ocean demersal fish stocks. It was agreed that the UNCLOS, the UNFSA, the FAO Compliance Agreement, and the Code of Conduct for Responsible Fisheries were the relevant instruments, particularly articles 116-119 of the UNCLOS. This issue has been developed further in meetings organised by the FAO. The downturn in northern-hemisphere demersal fisheries will inevitably cause a shift of effort to southern demersal fisheries in high seas waters, heightening the need for management controls. Whereas the Tasman fisheries based on the two straddling stocks mentioned above have been subject to catch controls, this form of management was only made possible by an agreement between Australia and NZ. The southern Indian and Southern Oceans may well contain other straddling stocks of significant commercial importance, including species other than orange roughy and alfonsino. Information on straddling and high seas stocks is needed to assist management policy development. Even in a current absence of fishing, the known spatial and depth distributions of major demersal species can be used together with bathymetric and oceanographic data to postulate where straddling and high seas stocks are likely to occur. For example, the known distribution of orange roughy off Namibia and the bathymetry of the continental slope off southern Africa suggest that discreet straddling stocks of orange roughy may occur in that region.

Figure 1. Map of major orange roughy fisheries in the Tasman Sea area.

Information on the known and probable distributions of demersal fisheries in international waters is required for the development of regional management policies for such fisheries. Such information is fundamental to the development of effective regional management policies for these resources and could also prove useful to exploratory ventures by the Australian fishing industry. If Australia wishes to negotiate with another nation concerning controlled access to an AFZ straddling stock, a prior knowledge of what that nation has to offer by way of reciprocal exchange of access into their EEZ would greatly assist such negotiations.

The aim of this desk-top study was to establish an information base for major demersal resources in the high-seas areas of the southern Indian and Southern Oceans and Tasman Sea. Specifically the study aimed to:

• collate available information on major (non-shelf) demersal fish resources of the southern Indian and Southern Oceans;
• map the known and possible distributions of such resources; and
• delineate known and possible straddling stocks.

Data Review

Fisheries data in general

Commercial fishing is a highly competitive business in which catch data, particularly location data, are usually zealously guarded from prospective competitors. Each fisher typically strives to catch as large a share of a resource as possible before other fishers ‘discover’ it. Whereas the need for such confidentiality may lessen in established fisheries with well-known grounds, this need is heightened in new and developing fisheries, particularly when other nations are involved. For example, there are anecdotal reports of crew-members being offered A$100 000 to divulge the location of orange roughy “hot-spots” on the Madagascar Ridge during the early stages of this fishery. Hence, most fisheries databases are subject to strict commercial-in-confidence requirements that must be observed during scientific analysis. Individual shot data provide the most valuable information. Any tabling or publication of analytical findings is usually subject to predetermined data aggregation and spatial resolution limit criteria.

Ecological and biological data

Within the study area commercial fishing activities currently target three major species groups. To the far south, the major target species group contains toothfish and icefish. However, as the major fisheries for these species are south of the study area, the toothfish group was not considered for the current study. The other two major target species groups can be termed the orange roughy group and the alfonsino group. These two groups can occur in close proximity to each other.

Orange roughy group: Commercial species associated with orange roughy include black oreo (Allocyttus niger), spiky oreo (Neocyttus rhomboidalis), smooth oreo (Pseudocyttus maculatus) and deepwater sharks. In most instances these species are caught as a bycatch of targeting orange roughy. Orange roughy have a worldwide distribution (Kotylar 1996), but the bulk of the commercial catch has been taken from the southern hemisphere. They have been recorded at depths between 500 and 1800 m (Merret & Wheeler 1983) but are most abundant at depths between 750 and 1100 m where they often form large aggregations. Catch data within the Australian EEZ show abundance to peak between 850-950 m depth. However, orange roughy fisheries are known to occur at significantly shallower depths off Namibia. Aggregations are usually associated with bottom features such as seamounts and are often highly seasonal with spawning aggregations forming during the (southern) winter months of July and August. Although aggregations may extend up to 100 m off the bottom, orange roughy typically exhibit a “dive-flight” response to approaching fishing gear and a bottom trawl net is needed to catch them.

Orange roughy is a long-lived species with a maximum age of over 100 years and age at maturity of 25 to 35 years. In Australia, natural mortality is estimated to be between 0.057 and 0.068 depending on which sets of age composition data are used. New Zealand orange roughy assessments use a natural mortality estimate of 0.045. Spawning sites are known off eastern Tasmania (St. Helens seamount and off St Patricks Head) and on the Cascade Plateau and the South Tasman Rise. Spawning occurs on the Cascade Plateau in early June, off Eastern Tasmania through July and early August and on the South Tasman Rise (to date) from late July to mid-August. Spawning has also been recorded on some seamounts south of Tasmania and in the Great Australian Bight.

Orange roughy have comparatively low fecundity. Stock structure is briefly discussed later in this report.

Alfonsino group: Commercial species associated with alfonsino include boarfish, cardinalfish (Epigonus telescopus) and blue-eye trevalla or bluenose (Hyperoglyphe antarctica). Alfonsino also have a worldwide distribution, but are more wide ranging than orange roughy, occurring throughout tropical and temperate waters. Alfonsino have a maximum recorded age of 17 years and attain a maximum size of about 60 cm length, with females growing faster than males (Annala et al. 1999). Maturity is thought to occur at around 30 cm and 4 to 5 years age. Spawning grounds are unknown, but juveniles have been recorded in pelagic and epipelagic zones in the Indian and North Pacific Oceans. Although alfonsino have been recorded at depths between 25 and 1200 m, most commercial catches are made between 300 and 600 m. Trawl surveys on the Chatham Rise recorded alfonsino between 250 and 650 m depth (Bull et al. 2001). Alfonsino are also typically associated with seamounts and similar features but often school in the water column and can be caught with midwater trawl nets.

Because of the narrower water temperature and depth ranges for orange roughy when compared with those for alfonsino, orange roughy was selected as the most suitable species for this initial attempt at modelling distribution.

International and high seas demersal fisheries data

Up until recently, the limited high-seas demersal fishing activities by the Australian fleet have been generally poorly monitored. Whereas some historic Australian shot-by-shot data are on file on AFMA logbook databases, these data were supplied on a voluntary basis. It is now a statutory requirement for Australian trawlers obtaining permission to fish high seas waters to complete shot-by-shot logbooks and have a (satellite) vessel monitoring system, as well as carrying onboard observers when required. The catch-recording situation was very similar for New Zealand vessels fishing the high seas, but recent legislation has since made it a statutory requirement to supply shot-by-shot data.

Australian fishing companies that had submitted high seas logbook data were contacted directly regarding the use of their data in this project. All were in agreement to the data being used, provided confidentiality was maintained. However, because of the competition between Australia and other nations for high seas demersal resources in the study area, it did not prove possible to acquire shot data from other sources. Under a Memorandum of Understanding with Australia, New Zealand provides shot data for their vessels fishing the South Tasman Rise orange roughy fishery, but no other high-seas data were available.

Several nations are now involved in targeting demersal resources in Southern and Indian Ocean high seas waters. Also, many vessels are flying flags of convenience. A FAO meeting held in La Réunion in February 2001 to facilitate the establishment of a South West Indian Ocean Commission (SWIOC) noted that there was an urgent need to establish a reliable catch database for such waters. Consequently, the FAO convened an Ad hoc scientific Meeting on the Management of Deepwater Fisheries Resources of the Southern Indian Ocean in Swakopmund, Namibia on 30 May to 1 June 2001. At this meeting it was agreed that all countries should initially provide catch (area, species and weight) and effort (at least numbers of vessels and trawl shots) data to the FAO secretariat which would handle these data in accordance with an agreed protocol on confidentiality. A protocol for future reporting of data on a tow-by-tow basis was also drafted. It was recognised that some countries collect more detailed information than that prescribed in the data collection protocol. For this reason, it was agreed that if national authorities were collecting appropriate data, their vessels should continue to follow their statutory or regulatory requirements. However, countries that had no formal provisions for data collection for vessels operating on the high seas should follow the meeting’s data collection requirements.

A draft copy of the meeting’s discussions and policies on data treatment is appended as Appendix II.

The meeting further agreed that until a future fisheries commission provided management direction for the area, individual national authorities would be the repositories for data recording the activities of their flag vessels in the study area. The (FAO) secretary offered to contact national authorities in the case of countries where it was uncertain if the data was being appropriately archived to advise them of the future expected management process and data requirements. It is anticipated that catch data will be tabled for analysis at a May 2002 meeting of this scientific group. However, these data were not available for the current study.

AFMA logbook and CSIRO survey data

AFMA collects catch and effort data from Commonwealth fishers within the AFZ through the AFMA logbook program. Catch and effort data are stored in the AFMA logbook database, which was developed from the Australian Fishing Zone Information System (AFZIS). AFZIS was initially developed as a joint project between the (then) Australian Fisheries Service (AFS) and the CSIRO Division of Fisheries, and was intended to meet most of the AFS's data processing and storage needs, integrating fisheries catch and effort data, research data, and licensing and compliance records. However, major parts of the system were never implemented.

When the AFMA was formed it took over responsibility for AFZIS logbook databases. These databases are in the process being redeveloped. The existing system of developing a new database and associated data entry application based on a logbook design has resulted in a complex and fragmented approach to data collection. To overcome the problems associated with this approach, AFMA has undertaken a review of all existing and proposed logbooks. The work completed to date has resulted in a generic data model and dictionary, which are currently being reviewed by data users.

Submission of catch and effort information is usually a condition of a fishing permit or licence. Although data managers have the backing of legislation in acquiring data from fishers, it is cooperation and trust from industry that are the essential ingredients for efficient collection of high quality data. Three fisheries and associated logbook databases managed by AFMA were identified as containing significant numbers of orange roughy catch records by Australian demersal trawlers. These were the Western Deepwater Trawl Fishery (WDTF), the Great Australian Bight Fishery (GABF) and the South East Fishery (SEF). The latter database also includes catches from the South Tasman Rise (Figure 1).

The Western Deepwater Trawl Fishery logbook database (WDT01) holds catch and effort data by individual operation and has been in use since 1988. The database has records of approximately 2000 shots. The Great Australian Bight Fishery logbook database (GB03) also holds individual shot data, has been in use since 1977, and has records of approximately 40,000 shots. The South East Fishery logbook database (SEF1B) again holds individual shot data, has been in use since 1985, and contains approximately 550,000 shots.

Fishing operations from the Indian and Southern Oceans recorded in the Great Australian Bight and the Western Deepwater Trawl fishery logbooks were considered first by extracting all records with longitudes of less than 100 degrees East. There were 260 species records from 130 geographically unique operation locations dating back to 1997. A total of seven species were recorded from these operations including whiptails, mackerel icefish, boarfish, alfonsino, orange roughy, oilfish and Patagonian toothfish. Searches for these species were then made on the entire Western Deepwater Trawl fishery database, Great Australian Bight fishery database and South East Fishery database. This revealed 637 records dating from 1993 in the WDTF database (307 geographically unique operation locations), 10562 records dating from 1986 in the GABF database (2816 geographically unique operation locations) and 83225 records dating from 1985 in the SEF database (7903 geographically unique operation locations). A further 147 records from a Bureau of Rural Resources coordinated multi-vessel trawl survey of the deep-sea demersal fisheries resources of the Great Australian Bight in 1988 were identified (Newton and Klaer 1991) in which orange roughy were caught at 35 sites.

CSIRO and some State fisheries agencies have also conducted fish surveys in Australian waters, including occasional deepwater trawl surveys. One such survey conducted on the continental slope of Western Australia in 1991 is described in a report to FRDC (CSIRO 1992). Minimum and maximums for depth, latitude and longitude are reported by species, but unfortunately species are not specifically assigned to trawl survey locations in the report. Several other deepwater trawl surveys have been carried out over the last decade, generally in waters off Tasmania and on the continental slope of southern Australia, as well as off Western Australia and around Macquarie Island and the Macquarie Ridge. Whereas these datasets are accessible by arrangement with the individual scientists involved and/or the relevant custodian(s), the time constraints associated with this desktop study precluded their use.

Oceanographic data

Internet searches for oceanographic information revealed enormous repositories of spatial data for environmental factors such as temperature, salinity, chlorophyll, nutrients, physical and chemical properties, sea surface height, wind stress and other parameters. However, much of the information consisted of raw data from shipboard monitoring, drifting buoys or satellite sweeps. Relatively little data have been interpolated to grids of global extent, or have been averaged or otherwise summarised to long-term cyclical temporal scales such as seasons or years. Where these types of spatial and temporal analyses had occurred, the subsequent data were usually only available at a spatial resolution of one degree, or at best at 15 nautical miles (nm).

Global bathymetric and sea surface temperature grids were also available from within the BRS’s own data archives. The bathymetry data existed in raster (grid) format with a spatial resolution of 2 nm and a horizontal resolution of 1 to 12 km. The grid was derived by combining available depth soundings with high-resolution marine gravity information from the Geosat and ERS-1 spacecraft. Previous global bathymetric maps lacked features such as the 1600-kilometer-long Foundation Seamounts chain in the South Pacific. The temperature data existed as monthly climatology summaries of sea surface temperature (SST) data using the V4, V4.1, and interim V4.1 NOAA/NASA AVHRR Oceans Pathfinder data. The climatology summaries were created using daily averaged SST over monthly periods for 1985-1997 at 9.28km resolution. The data were originally provided to the Physical Oceanography Distributed Active Archive Centre by Dr. Ken Casey at the NASA Goddard Space Flight Centre.

Spatial Modelling of Orange Roughy Stocks

Methods

The study area was defined as the seas and oceans within a box bounded by the coordinates 0°E, 0°S and 150°E, 55°S, stretching from the southeast Atlantic Ocean to the Tasman Sea (Figures 2 and 3). This area includes the original SWIOFC Agreement area (Figure 3) that will now probably be expanded eastwards to the Australian EEZ (Figure 2) to about 120⁰E (the SWIOFC will now probably become the Southern Indian Ocean Fisheries Commission). The area also includes northern parts of the Commission for the Conservation of Marine Living Resources (CCAMLR) management region(s). The bathymetric features of significance in the study area are named in Figures 2 and 3.

Records of orange roughy catches from vessels operating in the Great Australian Bight Fishery, the Western Deepwater Trawl Fishery and the South East Fishery were extracted from AFMA databases. Four geo-referenced records of orange roughy ‘hot spots’ from the Namibian fishery were also included.

Shots (records) falling outside the study area and shots containing less than 1000kg of orange roughy were initially removed from the dataset. Spatially imprecise records and error records were also removed by intersecting fishing points with a bathymetry map and deleting all records occurring at depths greater than 2000m.

Two approaches to predicting the distribution of orange roughy stocks were used in this study. The first approach involves simply identifying the geographic distribution of locations with the same, or similar, environmental conditions to those at sites where the species has been caught. The use of general-purpose geographic information systems to build niche distribution models for species or assemblages in this way is generally known as subjective empirical modelling. The technique has more commonly been developed for predicting the distribution of terrestrial biota, but is equally applicable to marine species provided there are sufficient oceanographic, biological and catch data for the species of interest.

Statistical modelling methods were also investigated, namely generalised linear modelling and generalised additive modelling. Generalised linear modelling is an extension of ordinary linear regression. Linear regression fits linear (straight line) functions relating a response (dependent) variable to one or more predictor (independent) variables. A basic assumption of linear regression is that the relationship between the response variable and each of the predictors can be approximated by a straight line. Generalised linear modelling removes this assumption by providing a class of models that allow nonlinearity in response functions. Generalised additive modelling is an extension of generalised linear modelling that relaxes previous assumptions concerning the functional form of species' responses to environmental variables. The principal difference between generalised additive models and generalised linear models in modelling species distributions is that generalised additive models allow the survey data to determine the shape of response curves, instead of being constrained by specified parametric forms. In other words, fewer assumptions are made about how species respond to their environment.

Both generalised linear modelling and generalised additive modelling have been used to model species distributions, most commonly as a logistic regression with a binomial (presence versus absence) response. These techniques generate a probability that a species will be present in a given location defined in terms of its environmental attributes.

Figure 2. Bathymetric features of the eastern half of the study area

Figure 3. Bathymetric features of the western half of the study area

The ideal site-based data set for spatially modelling biological data contains population numbers of all species from an unbiased survey of geo-referenced sites and includes measurements of all relevant environmental variables (Belbin et al. 1995). Austin and Heyligers (1989, 1991) and Austin and Meyers (1995) argue that representative sampling of the environments and geography of a region is an appropriate strategy for most ecological purposes, including modelling habitat and species distributions. In the marine environment therefore, surveys designed using an environmental stratification, including geographic replication where possible, and with sites systematically or randomly sampling bathymetric variation, should potentially provide the best data for modelling spatial patterns.

The available orange roughy data were obviously biased by the targeted rather than representative or random nature of fishing operations and by the sampling capacity of the fishing gear. The maximum depth for demersal trawling shots is currently around 1500 m and therefore the relative abundance of fish stocks at greater depth is largely unknown. However, available catch data for orange roughy suggest that depths >1500 m are sub-optimal for this species. Spatial models of the likelihood of occurrence of the species can still be developed, but should always be interpreted cautiously with regard to the quality of both the logbook and environmental data and the known and potential sampling bias.

Data analysis

Subjective empirical modelling

Depth and sea surface temperature were chosen as the predictive variables for describing the distribution of orange roughy using subjective empirical modelling techniques because of their known (direct or surrogate) influence on fish distributions and the relative quality and availability of data compared with other oceanographic variables.

The proportion of orange roughy catch and catch-per-unit-effort was examined with respect to depth using trawl depth data recorded with shot records from the Western Deepwater Trawl (1993-2000), Great Australian Bight (1993-2000) and South East (1999-2000) fishery logbooks and available Australian data from the Southwest Indian Ridge (1999). These data were summarised by 100m depth strata (Table 2) then CPUE was further aggregated into 200m depth strata (Figure 4). A function representing the relative likelihood of occurrence of orange roughy in each depth class, ie, the estimated proportion of total fish biomass expected in each depth class, was subsequently developed on the basis of this analysis and expert input from fisheries scientists in the Bureau.

Aggregating/analysing all the available catch by depth data for orange roughy would have resulted in the large South East Fishery (SEF) database dominating (and probably skewing) data from elsewhere. SEF catch and effort by depth data for orange roughy are summarised in Table 1. Because the bulk of the catch was taken from seamounts, such as St Helens Hill off eastern Tasmania, most fishing effort was expended in the 800-999 m depth range and catch rates were highest in this depth band. The high percentage of shots that contained orange roughy (>80%) in all recorded depths over 800 m (Table 1) also illustrated the high degree of targeting accuracy in this fishery that accompanied the fishing of “known” orange roughy grounds/features. It should also be remembered that other fish species of commercial value at these depths are essentially limited to the lower-value oreos and sharks, which makes orange roughy the prime target species.

Table 1. Summary of SEF orange roughy catch by depth data 1986 to 1998, inclusive.

Table 2. Catch and effort data for orange roughy by 100m depth strata from the Western Deepwater Trawl, Great Australian Bight and South East fishery logbooks and FV ‘Austral Leader’ data from the Southwest Indian Ridge.

Figure 4. Orange roughy CPUE by depth strata using data from the Western Deepwater Trawl, Great Australian Bight and South East fishery logbooks and FV ‘Austral Leader’ data from the Southwest Indian Ridge.

The potentially disproportionate representation from using the entire SEF database was avoided by using 1999 and 2000 data only, noting that in these years the database also encompassed significant fishing effort on the Cascade Plateau and South Tasman Rise.

To examine the proportion of orange roughy catch and catch-per-unit-effort with respect to sea surface temperature, catch records from the Western Deepwater Trawl (1993-2000), Great Australian Bight (1993-2000) and South East (1999-2000) fishery logbooks and a FV Austral Leader log-sheet from the Southwest Indian Ridge (1999) were intersected with a mean annual sea surface temperature grid. Catch and effort data were then summarised by 2°C temperature strata. A function representing the relative likelihood of occurrence of orange roughy in each temperature class, ie, the estimated proportion of total fish biomass expected in each temperature class, was subsequently developed on the basis of this analysis and expert input from BRS fisheries scientists.

To generate a spatial (mapped) result, the intervals and associated relative probabilities of occurrence for the depth and temperature functions were used to reclassify the bathymetry grid, and seasonal (summer, autumn, winter, spring) sea surface temperature grids. Data in the reclassified grids were then added together. The resulting seasonal prediction grids contained relative likelihood of occurrence estimates that were the sum of the probabilities from the classified bathymetry and seasonal sea surface temperature grids. The prediction grids were subsequently bounded with respect to latitude. The most northerly record of orange roughy catch was off the Namibian coast, approximately 19º S. This record was used to define the most northerly extension of the species’ range and predictions north of 19º S were down-weighted to 0.

Statistical modelling

Random or representative demersal trawl survey data covering the geographic space and environmental gradients of the study area were not available. Consequently, there were no appropriate absence data with which to rigorously model the distribution of orange roughy using statistical techniques. Thus, the orange roughy catch data were effectively ‘presence only’ data. To circumvent this problem and allow the application of logistic regression techniques, the data were enhanced by the inclusion of approximately 1000 ‘pseudo-absences’ spatially randomly selected from the study area. Generalised linear and generalised additive models were fitted with a weighting applied to each of the pseudo-absence sites to emulate an equal number of presences and absences. This was achieved using the case weighting option provided by S-PLUS (Hastie 1992). Each pseudo absence site was given a weight of nP/nA where ‘nP’ was the number of presence records for the species and ‘nA’ the total number of pseudo absences. Each presence site was given a weight of one. This weighting facilitated estimation of approximate degrees of freedom, deviances and significance levels appropriate to a presence-only model. The weighting also enabled predictions to be expressed in terms of an index of relative likelihood of occurrence ranging from zero to one. Unlike predictions derived from presence versus absence modelling this is only a relative index, not an estimate of probability. Confidence limits cannot be derived for presence-only models due to the weightings and approximations applied in deriving these models (see NSW NPWS 1994 for more detail on this modelling methodology).

Orange roughy records were intersected with a bathymetry grid and sea surface temperature grids for the months January, April, July and October. The records were also intersected with a 30 nm (½ degree) grid of salinity at a depth of 1500m generated from the 1994 World Ocean Atlas (Boyer and Levitus 1997). Generalised linear and generalised additive modelling techniques were used to model the distribution of the species as a logistic regression. A forward and backward stepwise procedure was used to achieve the best possible model from those predictive variables found to have a significant relationship with the orange roughy catch data.

Once a model had been created, the likelihood of occurrence of orange roughy was spatially interpolated by applying the fitted model to the environmental grids within the study area. A look-up table containing function values for the fitted functions was derived and exported from S-Plus. This table contained fitted function values for 100 values of each predictor, evenly spaced across the range of the predictor. The look-up table was used in conjunction with the grids for the environmental predictors featuring in the model to predict the likelihood of the species occurring in each grid cell within the study area. The resolution of the grid cells in the final output grid was 5 nautical miles; the spatial resolution of the coarsest predictive environmental grid in the model, ie, the sea surface temperature grids.

Results

Records of orange roughy catches from AFMA datasets dated back as far as 1986, but were predominantly from the 1990s.

Subjective empirical model

The relative likelihood of occurrence function developed for orange roughy with respect to sea surface temperature is shown in Figure 5. This represents the estimated proportion of total fish biomass expected in each temperature class. The function developed with respect to depth is shown in Figure 6, representing the proportion of total fish biomass expected in each depth class.

Figure 5. Relative likelihood of occurrence of orange roughy with respect to sea surface temperature.

Figure 6. Relative likelihood of occurrence of orange roughy with respect to depth.

A simple model of the relative likelihood of occurrence of orange roughy, or index of relative abundance, was achieved by combining (adding) the temperature and depth functions. The index values ranged between 0 and 60 in the study area.

Model Predictions

Spatial predictions of the likelihood of occurrence of orange roughy stocks in the east Indian Ocean and along the southern Australian coast in each season based on the subjective empirical model are shown in Maps 2, 4, 6 and 8. In summer and autumn, the model predicted potentially significant stocks of orange roughy around Amsterdam Island in the central southern Indian Ocean, but predictions were only moderate for Broken Ridge and the East Indianman Ridge. The locations of these bathymetric features are shown in Figure 2. Likelihood of occurrence was relatively low along the west Australian coast, becoming moderate to high off southwest Western Australia and through the Great Australian Bight. The model predicted high to very high chances of finding orange roughy stocks south east of Kangaroo Island, around the south of Tasmania and north to Flinders Island. Predictions on the South Tasman rise were generally moderate to high.

In winter and spring, predictions of orange roughy presence on the Broken Ridge and the East Indianman Ridge, along the west Australian coast and in the Great Australian Bight increased. Predicted likelihood of occurrence also increased in northeast Bass Strait. Conversely, the predicted likelihood of occurrence of orange roughy declined in western Bass Strait and around Tasmania, on the South Tasman Rise and around Amsterdam Island.

Spatial predictions of the likelihood of occurrence of orange roughy stocks in the west Indian Ocean and around the southern African coast in each season based on the subjective empirical model are shown in Maps 3, 5, 7 and 9.

In summer and autumn, the subjective empirical model predicted the likelihood of occurrence of orange roughy stocks off the central and southern Namibian coast to be relatively high. In contrast, predictions on the upper continental slope off the west coast of South Africa were only low to moderate and low for the rest of the South African, Mozambique and Madagascar continental slope regions. The model predicted low to moderate chances of finding orange roughy stocks on the Madagascar Ridge. Relatively high likelihoods of occurrence were predicted for southern and central parts of the Southwest Indian Ridge. Moderate to very high predictions were also noted to the south west of South Africa on the Wyandot and Schmidt-Ott seamounts and the Cape Rise. The locations of these bathymetric features are shown in Figure 3.

Predictions of the likelihood of occurrence of orange roughy stocks in most of the aforementioned areas of the west Indian Ocean and African continental slope increased in winter and spring. Predictions of the presence of orange roughy also appeared in winter and spring on the Mozambique Plateau. However, the likelihood of occurrence of orange roughy in the southwestern parts of the Southwest Indian Ridge declined. As mentioned in the previous section, the predicted presence of orange roughy around Amsterdam Island also declined in winter and spring.

Statistical model

Following the data checking and filtering process described in the methods section, there remained 675 records of catches of orange roughy in the study area that were used in the statistical modelling analysis. This represented 465 geographically unique sites based on the grid cell size used in the spatial interpolation process, namely 5 nautical miles, the resolution of the SST grids. The distribution of Australian EEZ and South Tasman Rise catch records used in the statistical modelling process is shown in Map 1. The catch data were mostly concentrated along the upper continental slope between Kangaroo Island and the north east coast of Tasmania. Orange roughy catches in this area comprised 71% of the total modelling dataset. Catches in the waters off the south coast of Tasmanian alone comprised 40% of the dataset. There were 58 records (12%) from the area around the St Helens sea mount, 23 records (5%) from the South Tasman Rise and 158 (34%) from the Great Australian Bight and Western Australia. Analyses also included 23 (5%) records from the Indian Ocean, but these are not shown on Map 1 for reasons of commercial confidentiality.

No significant relationship was found between orange roughy catch and salinity data from the ¼ degree salinity grid. Significant relationships however, were found with bathymetry and with all of the sea surface temperature variables. A generalised additive model based on bathymetry and April sea surface temperature was ascertained to be the best statistical model describing the distribution of orange roughy. This conclusion was based on both the amount of deviance explained by the model as well as expert evaluation of the mapped predictions of likelihood of occurrence for the species. The model results are shown in Figure 7.

Orange roughy

All records 465 Total pseudo sites 1466
Null deviance 1289.25 on 1465 df
Residual dev. 130.64 on 1461 df
Deviance explained 89.87 % Model type: GAM

Figure 7. Statistical modelling results for the generalised additive model based on bathymetry

Figure 7. Statistical modelling results for the generalised additive model based on April sea surface temperature.

Model Predictions

Spatial predictions of the likelihood of occurrence of orange roughy stocks in the east Indian Ocean and along the southern Australian coast based on the statistical model are shown in Map 10. The statistical model (generalised additive model) - using the variables ‘bathymetry’ and ‘April sea surface temperature’ - predicted the likelihood of occurrence of orange roughy on the Broken Ridge and the south-western coast of Western Australia to be moderate to high, relative to areas such as the Ninetyeast Ridge. This is a similar result to that described by the subjective empirical model using Depth and Autumn Sea Surface Temperature, although predictions by the statistical model off parts of Western Australia and on the East Indianman Ridge are inflated. The statistical model predictions were moderate to high around Tasmania and around Amsterdam Island. The subjective empirical model based on Autumn Sea Surface Temperature returned high to very high predictions in these areas relative to other parts of the distribution.

Spatial predictions of the likelihood of occurrence of orange roughy stocks in the west Indian Ocean and around the southern African coast based on the statistical model are shown in Map 11. The statistical model predicted moderate to high chances of locating orange roughy stocks off the central and southern Namibian coast. Unlike the subjective empirical model based on depth and Autumn sea surface temperature, the statistical model showed quite a high likelihood of occurrence of orange roughy stocks off the southern South African coast. The model predicted moderate to high chances of finding orange roughy stocks on the Madagascar Ridge and Southwest Indian Ridge and high to very high likelihoods of occurrence of stocks on the Mozambique Plateau, the Del Cano Rise and to the north of the Crozet Islands and Kerguelen Island.

Discussion

Model Performance

It was intended that a contoured representation of the relative intensity of orange roughy catch throughout the study area be attempted by applying a spatial neighbourhood based density analysis to geo-referenced catch data using ArcView Spatial Analyst tools. However, the paucity of high-seas catch data able to be collated to date rendered this analytical technique unsuitable in most parts of the study area, except perhaps in south eastern Australia. In this region, enough fishing has occurred to generate catch and effort contours from geo-referenced logbook records and to spatially represent change in catch over time. To some extent, this type of analysis of demersal fishing data in south eastern Australian waters has already been used in stock assessments for orange roughy.

The subjective empirical modelling approach appeared to predict reasonably well on a regional scale when compared with AFMA logbook records for the Great Australian Bight, Western Deepwater and South East trawl fisheries. Predictions were generally low along the West Australian coast where Western Deepwater Trawl logbook records and independent surveys conducted by CSIRO have indicated little in the way of orange roughy stocks. In comparison, the statistical model appeared to be over-predicting along the West Australian coast based on this information. The statistical model also appeared to be under-predicting around Tasmania, where the majority of significant orange roughy catches have been recorded. This is probably due to the model being unable to adequately account for seasonal effects. Early catch, effort and CPUE data suggest the species was most prevalent in this area over the summer months, though in recent years most catches have been made during the winter spawning season.

Both the subjective empirical models and the statistical model predicted quite well with respect to the known orange roughy hotspots off the Namibian coast. The subjective empirical models derived using summer and autumn sea surface temperature also matched well with known catches from the southwest Indian Ocean and predicted low likelihoods of occurrence off the southern South African coast where exploratory fishing for orange roughy has so far failed to uncover any significant stocks. In contrast, the statistical model and the subjective empirical models derived using winter and spring sea surface temperature showed quite a high likelihood of occurrence of orange roughy stocks off the southern South African coast. Available data and industry opinion suggests these models are over-predicting in this area.

The statistical model predicted moderate to high chances of finding orange roughy stocks on the Madagascar Ridge and Southwest Indian Ridge where stocks have been located. However, high to very high likelihoods of occurrence of stocks on the Mozambique Plateau, the Del Cano Rise and to the north of the Crozet and Kerguelen Islands are currently not supported by known fishing activities or anecdotal fishing industry opinion.

The use of sea surface temperature as an indicator essentially assumes a more or less constant decline in temperature with depth. Obviously, this does not occur in many instances with thermal stratification, currents, gyres and upwellings having a pronounced influence on the distribution of temperature at depth. Orange roughy do not normally occur within 4-500 km of the Antarctic Convergence and it is probable that the high likelihoods of occurrence of stocks on the Del Cano Rise and to the north of the Crozet and Kerguelen Islands are driven by ‘warm’ surface water temperatures overriding the cold Antarctic water temperatures at orange roughy depths. There are unsubstantiated reports of illegal toothfish fishing having occurred on the Del Cano Rise that further suggests that this area is too cold for orange roughy.

The need for statistically rigorous methods of predictive modelling depends on the purpose of the modelling exercise and the choice of data and method depends on the predictive confidence required to satisfy the purpose. There is naturally a strong sampling bias associated with fisheries logbook data sets. Most records occur around known ‘hotspots’, limiting confidence in the results of modelling. Predictive methods capable of providing statistical confidence limits for their predictions require at least presence/absence data. The nature of the fisheries data available for analysis in this project are such that they should be used for predicting the potential distribution of species; that is, for purposes where it is better to include areas without the species than exclude areas with the species. The subjective empirical models developed require further expert and empirical evaluation before they can be used on a routine basis.

Access to fishing data from other countries is being negotiated via the FAO, albeit that the commercial confidentiality aspects of these data currently renders them hard to obtain for new, ‘developing’ fisheries. More effort should be directed towards collating or deriving better quality oceanographic data. Data on current velocity and direction, water mixing, gyres, upwellings, temperatures at depth, nutrients, etc. may prove very useful for predictive spatial modelling of demersal stocks, but are currently not readily available at an appropriate spatial resolution on global or oceanic scales. In future studies, it is recommended that specific study areas and objectives be determined before attempting to gather such data.

Orange Roughy Stock Structure and Behaviour

The stock structure of orange roughy in the southern Indian and Southern Oceans is unknown. However, the information obtained from stock structure studies in the Tasman Sea and Australian and New Zealand EEZs, indicates that it is probable that there are several discrete stocks in these oceans. Although orange roughy are known to spawn in the water column and eggs and larvae can theoretically be quite widely advected, little is known of their early life history and movements. Most established fisheries appear to occupy distinct areas, usually seamounts, plateaus or ridges, with some stock structure studies on adult fish suggesting a high degree of fidelity to these areas.

No single stock structure method currently available is universally agreed upon as providing unambiguous evidence of stock structure in marine fish species. The accepted approach is to apply several independent techniques to a problem and test for similarity of outcomes as indicative of underlying ‘real’ stock structure. Genetic studies of orange roughy stock relationships have focused on allozyme frequencies and mitochondrial DNA distributions. These have been applied over a number of Australian and New Zealand fisheries. Results have been mixed, with concern that a low rate of mixing between populations can inhibit genetic divergence. However, mitochondrial DNA has clearly identified separate stocks of orange roughy on the Challenger Plateau, the Puysegur Bank, and the Chatham Rise, and between western and eastern regions of southern Australian waters. Allozyme and mitochondrial DNA work on Tasman Sea fish indicated differences between the Lord Howe Rise, Northwest Challenger Plateau, and Southwest Challenger Plateau. Otolith microchemistry analyses on orange roughy sampled at sites across Australian and New Zealand waters and the Tasman Sea indicated highly significant differences in the composition of otoliths of fish from different sites. These data suggested five stocks in Australian waters, possibly two stocks in New Zealand waters, and independent stocks on the Lord Howe and South Tasman Rises. Conversely, a recent microsatellite DNA study (Oke et al. 2000) which looked at fish over a wide area (including Namibia and the Atlantic Ocean) failed to find discrete population groupings, suggesting that migrants from several populations may sustain or re-establish orange roughy stocks. However, a weak but significant isolation by distance effect was also detected.

Whereas these independent studies often have different findings, the overall balance of evidence suggests that a significant proportion of adult orange roughy eventually take up residence on or around underwater landforms such as seamounts, plateaus or ridges. Although some genetic exchange may occur between geographically isolated populations, each population should be managed as a discrete stock unless there is evidence of significant mixing between adult populations.

Assessment of orange roughy stock structure and prediction of distribution are also complicated by seasonal migrations. Off Tasmania, otolith shape analysis data indicate that spawning eastern zone fish and southern zone non-spawning fish may comprise a common stock, which is distinct from an eastern non-spawning and southern winter-caught “stock”. Thus, the ‘St Helens’ spawning hill draws fish from the south during winter. In New Zealand waters, spawning fish are known to travel considerable distances on the Chatham Rise. It is also probable that non-spawning migrations occur, possibly in association with seasonal changes in oceanographic parameters. Such movements may in part explain why the predictions of orange roughy occurrence in the northern half of the study area were generally greater in winter/spring, whereas in the southern half they were generally greater in summer/autumn. However, there is also the potential that the level and type of orange roughy response to different environmental factors may differ from year to year, and also that environmental variables of greatest importance cannot be readily derived at an appropriate spatial resolution throughout the area of interest.

Interpretation of fisheries data to deduce comparative seasonal abundance (or presence) is further complicated by there often being ‘resident’ and ‘migratory’ populations within a common orange roughy stock. Residents often form ‘non-spawning’ aggregations, whereas the dispersed migratory fish are only accessible to fishing when they aggregate during the spawning season. With the development of a fishery, the resident fish are usually targeted throughout the year and are consequently fished-down at a greater rate than the migratory fish. Thus, in a “developed” fishery the bulk of the catch typically occurs during the winter spawning season, with little fishing effort occurring at other times of the year. Consequently, there is often a seasonal bias driven by the state of development of an orange roughy fishery. Interpretation requires a comprehensive catch history of the fishery.

Obtaining knowledge of stock structure and fish movement requires significant research expenditure. In the absence of such information, the most practical “stop-gap” and precautionary approach from a management perspective, is to initially assume that each significant “isolated” geographic feature (such as an ocean ridge) on which orange roughy occur supports a distinct stock. Orange roughy typically occur at distinct sites, such as seamounts, along/on such features. Defining what distance constitutes “isolated’ is problematic, but it should probably be at least 2-300 km. As noted above, in the Australian EEZ orange roughy travel from south of Tasmania to the St Helens spawning aggregation, a distance of over 150 km. However, the Cascade Plateau, a geographic feature lying about 200 km southwest of St Helens is known to have an orange roughy population distinct from those around Tasmania. Also, the South Tasman Rise a geographic feature that straddles the Australian EEZ about 350 km south of Tasmania supports a distinct orange roughy stock. In the absence of fisheries data, predictive modelling should prove to be a useful tool in determining possible links between such geographic features. Fishing companies should also be requested to routinely collect size (length) and sex data and associated otoliths from the fish caught on these features.

Straddling stocks

No additional straddling stocks of orange roughy were clearly delineated by this study. Whereas the predictive modelling suggested straddling stocks may occur across the EEZs of the Le Crozet, Amsterdam and associated Islands and possibly off the South African coast more data are needed to assess the veracity of these suggestions. Detailed stock structure studies, such as those conducted on the South Tasman Rise orange roughy population, would also be needed for such a stock.

Conclusions

As the data improve, statistical models of fish abundance, or indices of abundance, derived using fishing effort data and oceanographic environmental data attributed to individual fishing records will become increasingly accurate and useful. However, the simplicity and flexibility of subjective empirical models is appealing. A workshop of experts convened to develop a series of response curves to various predictive oceanographic variables may yield interesting and valuable results.

More sophisticated data analyses may be possible as better fishing and environmental data become available. For example, fisheries data can and should be broken down by year and month and then modelled against oceanographic data such as sea temperature and current data that have also been derived by year and month. A FRRF funded project - Mapping standardised stock abundance indices for use in fisheries access, resource allocation and marine reserve planning and management processes addressing these issues is currently underway. Also, environmental data (depth temperatures, salinity, currents, sea floor slope and aspect etc.) collected at the time of fishing will generate more accurate ecological models than data attached to catch records by intersecting fishing operation records with spatially interpolated environmental grids.

The primary factors currently limiting the utility of spatial models of demersal fish resources are:

• the inherent bias of fisheries-dependent data,
• the poor geographic accuracy and precision of some of the fishing records,
• the unknown suitability of using parameters such as sea surface temperature for mapping demersal fish distribution,
• the low spatial resolution of much of the readily available oceanographic data, and
• the limited understanding of the movements and behaviour of demersal stocks such as orange roughy in response to environmental factors and fishing.

It is anticipated that a significant amount of high-seas fisheries data will become available via an AFFA/FAO scientific workshop on demersal fisheries in the study area, to be held in Perth in May 2002. These data should enable better ground-truthing of the current orange roughy distribution models and also facilitate broad-scale mapping of the alfonsino group.

Recommendations

1. If fishing operators are supportive, some form of iterative feedback process should be established between spatial modelling projects in BRS and trawling operations conducted to test and refine model predictions. It is also recommended that access to quantitative (presence/absence or count) data sets held by high seas fishing nations should be negotiated as soon as practicable.

2. Orange roughy data from AFMA logbooks and other sources should be used as an input data set in the FRRF funded project:- Mapping standardised stock abundance indices for use in fisheries access, resource allocation and marine reserve planning and management processes. The results from this research should then be fed back into any ongoing Indian Ocean demersal fish stock resource assessment work and compared with the models developed in the current project.

3. A workshop of experts on demersal fish stocks in the Indian and Southern Oceans, possibly in conjunction with the planned international deep sea fisheries symposium in New Zealand in 2003, aimed at better defining the environmental variables influencing stock distributions and abundance and the response of species to those variables, would be very useful in developing better spatial models of stock distributions.

4. Subjective empirical models of the distribution of demersal fish stocks in the S(W)IOFC Agreement Area should be developed next for alfosino given the relative amount and availability of alfonsino catch data and the anticipated direction of fisheries development in the southern Indian Ocean.

5. Predictions of fish distribution or abundance made using models derived from commercial fisheries data should always be treated with caution where spatial bias in the distribution of fishing operations is apparent.

6. If spatial modelling of both demersal and pelagic fish resources for fisheries management and policy development is seen as important work by the Fisheries and Aquaculture Branch and the Bureau of Rural Sciences, a new project to generate annual and seasonal grids of oceanographic variables such as salinity, chlorophyll, dissolved oxygen, thermocline depth, temperature and current velocity and direction at best possible resolution should be developed.

References

Annala, J. H., Sullivan, K. J. and O’Brien, C. J. (Comps) (1999) Report from the Fishery Assessment Plenary, April 1999: stock assessments and yield estimates. 430 p. (Unpublished report held in NIWA library, Wellington)

Austin, M.P. and Heyligers, P.C. (1989). Vegetation survey design for conservation: gradsect sampling of forests in north-eastern New South Wales. Biol. Conserv., 50: 13-32.

Austin, M. P. and Heyligers, P. C. (1991). New approach to vegetation survey design: gradsect sampling. In. C.R. Margules and M.P. Austin (eds.) Nature Conservation: Cost Effective Biological Surveys and Data Analysis. CSIRO, Australia: 31-36.

Austin and Meyers (1995). Modelling of landscape patterns and processes using biological data, Sub-project 4: Real data case study. Consultancy report to ERIN. CSIRO Division of Wildlife and Ecology, Canberra.

Belbin, L., Austin, M. P., Margules, C. R., Cresswell, I. D. and Thackway, R. (1994). Modelling of landscape patterns and processes using biological data, Sub-project 1: Data suitability. Consultancy report to ERIN. CSIRO Division of Wildlife and Ecology, Canberra.

Boyer, T.P. and S. Levitus (1997). Objective analyses of temperature and salinity for the world ocean on a 1/4 degree grid. NOAA Atlas NESDIS 11 (1997)

Bull, B., Livingston, M. E., Hurst, R. and Bagley, N. (2001) Upper-slope fish communities on the Chatham Rise, New Zealand, 1992-99. N. Z. J. Mar. Freshwat. Res. 35: 795-815.

CSIRO (1992). The fisheries biology of deepwater crustacea and finfish on the continental slope of Western Australia. Final report to the Fisheries Research and Development Coorporation (FRDC 1988/74), Canberra.

Hastie, T. J. (1992). Generalised additive models. In J.M. Chambers and T.J. Hastie (eds) Statistical Models in S. pp 249-308. Wadsworth and Brooks, California.

Kotylar, A. N. (1996) Beryciform fishes of the world ocean. VNIRO Publishing, Moscow.

Merret, N. R. and Wheeler, A. (1983) The correct identification of two trachichthyid fishes (Pisces, Berycomorphi) from the slope fauna west of Britain with notes on the abundance and commercial importance of Hoplostethus atlanticus. J. Natural History 17: 569-573.

Newton, G., and Klaer, N. (1991). Deep-sea demersal fisheries resources of the Great Australian Bight: a multi-vessel trawl survey. Burea of Rural Resources Bulletin No. 10. AGPS, Canberra.

NSW NPWS (1994). Fauna of north-east NSW forests. North East Forests Biodiversity Study Report No. 3. Unpublished report, NSW National Parks and Wildlife Service.

Oke, C. S., Crozier, R. H. and Ward, R. D. (2000) Stock structure of Australian populations of orange roughy analysed using microsatellites. Final Report to FR&DC. Project No. 97/118.

Appendix I Maps of the predicted distribution of Orange Roughy

Indian Ocean Demersal Fish Resources

Map 1 - The distribution of catch records for Orange Roughy in the Australian Fishing Zone used in the statistical modelling process.

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries & Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 2 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and summer sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries & Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 3 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and summer sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries, Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 4 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and autumn sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries & Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 5 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and autumn sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries, Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 6 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and winter sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries & Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 7 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and winter sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries, Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 8 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and spring sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries & Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 9 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and spring sea surface temperature - subjective empirical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture, Fisheries, Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 10 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and sea surface temperature - statistical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture Fisheries & Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Map 11 - Predicted likelihood of occurrence of Orange Roughy based on bathymetry and sea surface temperature - statistical model

Map produced by the Bureau of Rural Sciences
Fisheries and Marine Sciences Program for the
Fisheries and Aquaculture Branch
Agriculture Fisheries Forestry - Australia
December 2001
Projection: Geographic
Fisheries Data Source: AFMA
Copyright Commonwealth of Australia 2001

Bureau of Rural Sciences

Meeting Document 02/10 - Population Biology of Some Commercial Deepwater Fish Species of the Southwest Indian Ocean - Report on Progress

Malcolm Clark

NIWA

SPECIES COVERED

It was decided at the Swakopmund meeting that this report would have limited scope until more became known of detailed catch composition. It is based on the main known commercial target/bycatch species taken in the deepwater fisheries:

Original composition proposed was:

Following examination of more recent data, I will now add bluenose Hyperoglyphe antarctica (BNS). This is in the shallow group as a common bycatch of alfonsino. The black cardinalfish fishery appears more associated with the orange roughy fishery than the alfonsino, although to an extent it is also a stand-alone fishery.

SPECIES ACCOUNTS

The intended structure of the report follows the format below:

1. SCIENTIFIC NOMENCLATURE AND CITATION

2. MORPHOLOGY

3. BRIEF NOTES ON CHARACTERISTICS OF THE SPECIES

3.1 Global, distribution map
3.2 Habitat, bathymetric features
3.3 Depth range, preference
3.4 Seasonal distribution, migrations

4. COMMERCIAL FISHERIES

5. SIZE STRUCTURE

6. MAXIMUM SIZE, WEIGHT, LENGTH FREQUENCY, SEX RATIOS...

7. REPRODUCTION

7.1 Spawning periods
7.2 Spawning types (synchronous batch spawners etc)
7.3 Fecundity
7.4 Egg development
7.5 Larval and early life history

8. MATURITY

9. AGE AND GROWTH

Growth rates

10. LONGEVITY

11. FEEDING

12. PARASITES

13. COMMERCIAL FISHERIES

13.1 Fishing areas and seasons
13.2 Fishing methods
13.3 Catches

14. FISHERIES MANAGEMENT

14.1 Biomass estimates
14.2 Fisheries parameters (exploitation rates/harvest levels)

Sections in preparation for orange roughy are given to demonstrate the type of information to be included, and to generate feedback as to how appropriate this is for the groups’ needs.

EXAMPLE OF A SPECIES ACCOUNT

ORANGE ROUGHY

I. SCIENTIFIC NOMENCLATURE AND CITATION

Hoplostethus atlanticus collett, 1889, Family Trachichthyidae (slimeheads)

II. MORPHOLOGY

The orange roughy is one of 55 species of the Trachichthyidae family or slime heads, characterized by mucus cavities on the head, a distinct spine at the preopercle angle, the pelvic fins with one normal spine and six or seven soft rays and the abdomen with a median ridge of scutes. The color of the body is red to dark orange. The opercular membrane is black. In contrast to many teleosts, their swim bladder is not filled with a gas or air, but a waxy substance.

III. DISTRIBUTION

Global, distribution map: Orange roughy occurs in temperate waters of the North and South Atlantic Ocean, southern Indian Ocean, and in the South Pacific. It is not known to occur in the northern Pacific.

Habitat, bathymetric features

Orange roughy live in a range of habitat, and occur in small numbers over large areas in the appropriate depth range, but frequently form aggregations on or near seamount features and canyon edges. They generally stay within 50-100 m of the seafloor, and do not undertake extensive vertical migrations.

Depth range, preference

The species occurs on the upper continental slope between -450 m and -1700 m, but mostly from -800 and -1200 m. Depth preference varies geographically, with the fish most abundant at 800-1000m in New Zealand and Australia, 500-700 m off Namibia, and over 1000 m in the north Atlantic.

Seasonal distribution, migration behaviour

Orange roughy are not believed to undertake large migrations. However, they do move several hundred kilometres in some areas to and from the spawning grounds (e.g., Chatham Rise fishery in New Zealand, from southern Tasmania to St Helens seamount in Australia).

Commercial fisheries

Orange roughy is a highly-valued commercial species, and major trawl fisheries have developed around New Zealand, off Tasmania south of Australia, Namibia, west of the British Isles, and most recently off Chile and in the southern Indian Ocean.

IV. SIZE STRUCTURE

The size of orange roughy differs between parts of the world. Fish in New Zealand and Australia are typically about 35 cm SL, those off Namibia around 25-30 cm, and in the North Atlantic about 50 cm SL. These differences are possibly related to the length of time prior to maturation, after which adult growth is believed to be very slow. Interestingly, the size frequency of the stocks in New Zealand and Australian waters has not changed substantially, despite heavy fishing pressure for between 5 and 15 years. Reasons for this are uncertain, but it might reflect that recent recruitment levels (in the last 10-20 years) have been relatively low over a large geographical area, and new recruits have not been entering the population in sufficient numbers to shift the size structure towards smaller fish.

V. REPRODUCTION - SPAWNING PERIODS

Spawning occurs only once each year in the winter months (June to August in the southern hemisphere, January-February in the north Atlantic). Spawning occurs in dense aggregations, and is often associated with seafloor features such as seamounts and canyons.

Spawning development

Orange roughy are single determinate batch spawners, with developing maturation of gonads from 5-6 months prior to spawning during winter months. Off New Zealand, Australia, and Namibia spawning takes place between mid-June and early August. This occurs in a number of separate spawning grounds, with 11 major sites known in the New Zealand region, 2 in southern Australia, and 4 off the Namibian coast. The location and timing of the formation of aggregations for spawning is very consistent between years. The actual timing of spawning may be related to shortening daylight length towards the middle of winter. Off the British Isles, spawning takes place from the end of January with a maximum in February and March.

Spawning generally lasts about 3 weeks. During this time there may be some turnover of spawning fish, but it is not thought to be substantial.

Fecundity

Depending on fish size, females have from 30 000 to 380 000 eggs, although typically 40 000 - 60 000. Fecundity varies between locations, and over time within a population. In Australia, fecundity increased as stock size declined in the early 1990s, possibly a compensatory response to fishing. It is likely that females do not spawn each year, as in many areas off New Zealand and Australia about half the large fish do not develop their gonads in any given year.

Egg development

Large eggs (2 - 3 mm in diameter) are released near the bottom, and ascend to near the surface as protein changes affects their buoyancy. They then sink to close to the bottom again before hatching after about 10 days.

Larval and early life history

Little is known about larval stages and distribution, but small juveniles are believed to recruit to near-bottom waters after 1 year.

Maturity

Orange roughy mature when over 20 years of age. Their otoliths have a marked “transition zone” in banding which is believed to be associated with first spawning (Francis & Horn 1997). This has been used to estimate mean age at the onset of maturity, which ranges from 23 to 29 years for various New Zealand fishing grounds, around 23 years off Namibia, and 29 years for Tasmanian fish.

VI. AGE AND GROWTH

Ageing of orange roughy has seen considerable scientific debate, and there is continuing disagreement of the true age of adult fish (see review by Tracey & Horn)

Growth rates

Orange roughy grow slowly. Validated ages have juveniles at lengths of around 3 cm, 5.5, 7 and 9 cm during the first 4 years respectively.

Longevity

In the Southwest Pacific, the growth of orange roughy is believed to be very slow. Fish do not mature until about 25 years of age, and the bigger fish might live to over 100 years. Growth and longevity of orange roughy is subject to keen scientific debate. Age has been estimated by otolith zone counts, measurement of Pb-Rn radioisotope ratios, daily otolith growth increments, and micro chemical composition. Results have varied, although most indicate high longevity. Maximum ages from the various techniques range from 15-20 years to several hundred years. Juvenile ages (to 5 years) for New Zealand fish based on otolith zone counts have been validated by modal length analysis, but adult ages have not.

VII. FEEDING

In the Southwest Pacific, orange roughy feed on shrimps, squids and fishes. Prey composition varies with area, size of fish, and depth distribution. They are benthopelagic feeders, probably taking their prey close to the bottom.

VIII. PARASITES

IX. COMMERCIAL FISHERIES - FISHING AREAS AND SEASONS

Commercial trawl fisheries developed in the late 1970s off New Zealand, and since then new fisheries have started in a number of countries around the world. The main fishing grounds occur in the Southern Pacific, the Western Indian Ocean, the Southeast Atlantic and in the North Atlantic.

Fishing methods

The fish are caught by bottom trawl gear, with some limited mid-water fishing.

Catches

New Zealand: The fishery developed in the late 1970s. Current catches (2002) about 14 000 tonnes. Main fishing areas are the Chatham Rise and Ritchie Banks, with smaller fisheries throughout the EEZ. Historically catches of 50-60 000 tonnes, with the Chatham Rise sustaining recorded landings of 30 000 tonnes throughout much of the 1980s.

Australia: The main Australian fishery dates from 1989, with the discovery of aggregations off St Helens seamount and southern Tasmania. Reported catches peaked at 41 000 tonnes in 1990. Other fishing grounds include the Cascade Plateau and Great Australian Bight. Catches inside the AFZ are around 3000 tonnes.

Tasman Sea, international waters: South Tasman Rise, Lord Howe Rise, NW Challenger Plateau, Louisville Ridge.

Namibia: There are 4 main grounds, 3 on the general slope off the west African coast, and one on a seamount at the abutment of the Walvis Ridge.Catches reached 15 500 tonnes in 1997, before decreasing catch limits restricted the fishery. Catches are around 1000 tonnes.

North Atlantic: The French dominated early years of the fishery, and catches in 1991 and 1992 totalled between 4000 and 5000 tonnes in ICES areas VI and VII. Catches in those areas are now around 1000 t/yr. Area XII including the Hatton Bank has been fished by Faroese and French vessels, with several hundred tonnes per year.

Chile: Good catches were taken on seamounts off the coast in 1999, with a catch of 800 tonnes. In 2000 a TAC of 1600 tonnes was set, and fully caught.

Southwest Indian Ocean: This fishery began in 1999, with 10 000 tonnes taken off the SWIO Ridge. A similar level of catch was reported in 2000, although effort increased markedly.

Fisheries management - Biomass estimates

New Zealand: 9 orange roughy stocks have had assessments carried out. Estimates of current biomass for these total about 240 000 tonnes, which is about 30% of the original unfished total of approximately 800 000 tonnes. The largest stock is on the northeast Chatham Rise, estimated at about 300 000 tonnes original biomass. (Annala et al. 2001)

Australia: Eastern/southern Tasmanian stock estimated at 100 000 to 150 000 tonnes. Current stock size between 10% and 20% of this. GAB, Cascade not known, much smaller. (Wayte & Bax 2001)

Namibia: Combined total estimated from the 4 stocks is around 50 000 tonnes. Depletion varies between the grounds, about 23 000 tonnes, but 17 000 tonnes of that is from 1 area. The main historic fishery of Johnies is estimated at 11% of virgin size. (Brandao & Butterworth 2002)

North Atlantic: Stock sizes largely unknown. One estimate for Area VI, now at levels of 25-30 % virgin biomass. (ICES 2001).

Management practices

New Zealand: Area TACs, ITQs proportional to TAC, sub-area catch levels, individual feature limits in some quota areas (e.g. seamount catch restrictions). In some areas target 50% virgin, others 30% virgin (MSY).
Australia: Area TACs, ITQ on Cascade. Target MSY, 30%.
Namibia: Overall TAC, area-specific catch levels, target 50% virgin.
North Atlantic: Precautionary management advice, no fixed quotas.
Chile: TAC.
SWIO: No management, but discussions in progress
Tasman Sea: STR TAC 2400 t, between AUS and NZ. Lord Howe, NW Challenger,
Louisville, no catch restrictions nor management.

Fisheries parameters (exploitation rates/harvest levels)

Slow growth and high longevity means that orange roughy is a relatively unproductive species. Long-term sustainable yields for the fisheries are much lower than for most commercial shelf species, and their recovery from over fishing may be slow.

Natural mortality (M) has been estimated from otolith age data in New Zealand at around 0.045 yr-1, 0.06 in Australia (St Helens) and about 0.055 in Namibia. The derivation of M differs between countries. However, all estimates are low, and much lower than almost all other commercial species exploited in shallower waters on the continental shelf. Yield estimates used in various orange roughy fisheries vary depending upon the management requirements of individual countries. As an example of this low productivity, in New Zealand the long-term constant catch yield is around 1.5% of virgin biomass, and long-term MAY (assuming constant fishing mortality of about 6% of the present biomass rather than catch) is about 2% of the virgin stock size.

They are also vulnerable to depletion by commercial fishing, because of their dense schooling behavior, and the predictable location and timing of spawning and feeding aggregations. In both New Zealand and Australia, stocks have been fished down rapidly, and quota levels have had to change dramatically in recent years as stock sizes declined and in some cases became over fished.

The application of stock assessment models, and management have had varied success, but many stocks have been overexploited, and their recovery is uncertain. Issues of recruitment, variable aggregation behaviour between years, the influence of environmental conditions, and effects of fishing on behaviour, are largely unknown. The period of fishing is still very short relative to longevity. A conservative management strategy, slow and controlled development, lower catch levels than most people think, are essential. Gut feelings useful.