THE TIMOR-LESTE COASTAL / MARINE HABITAT

MAPPING FOR TOURISM AND FISHERIES

DEVELOPMENT PROJECT

PROJECT NO 1

Marine & Coastal Habitat Mapping in Timor Leste (North Coast) Final Report

Boggs. G., Edyvane, K., de Carvalho, N., Penny, S., Rouwenhorst, J., Brocklehurst,

P., Cowie, I., Barreto, C., Amaral, A., Smit, N., Monteiro, J., Mau, R., Amaral, J.,

Fernandes, L.

June 2009

ii

Marine & Coastal Habitat Mapping in Timor Leste (North Coast) – Final Report

Date: June 2009 Acknowledgement This Project was funded by the Ministry of Agriculture & Fisheries, Government of Timor Leste and was undertaken by the following staff of the Ministry, in collaboration with researchers from Charles Darwin University (CDU) and the Northern Territory Government, Department of Natural Resources, Environment, the Arts and Sport (NRETAS):

Guy Boggs (CDU), Karen Edyvane (NRETAS), Narciso de Carvalho (MAF-Fisheries), Shane Penny (NRETAS), Juno Rouwenhorst (CDU), Peter Brocklehurst (NRETAS), Ian Cowie (NRETAS), Celestino Barreto de Cunha (MAF-Fisheries), Anselmo Lopes Amaral (MAF-Fisheries), Neil Smit (NRETAS), Jose Monteiro (MAF-Fisheries), Pedro Pinto (MAF-Protected Areas), Raimundo Mau (ALGIS), Joao Xavier Amaral (MAF-Fisheries), Lucas Fernandes (MAF-Fisheries).

Café e Floressta

Agricultura Pescas

Loro Matan

This project is a recognised project under the Arafura Timor Seas Experts Forum (ATSEF). Citation This document should be cited as: Boggs G, Edyvane K, de Carvalho N, Penny S, Rouwenhorst J, Brocklehurst P, Cowie I, Barreto C, Amaral A, Smit N, Monteiro J, Pinto P, Mau R, Amaral J, Fernandes L. (2009). Marine & Coastal Habitat Mapping in Timor Leste (North Coast) – Final Report. Ministry of Agriculture & Fisheries, Government of Timor Leste.

Copyright of Photographs Cover Photos: Main: Karen Edyvane Central photographs left to right: Shane Penny, Guy Boggs, Karen Edyvane, Karen Edyvane, Tony Ayling, Shane Penny

Page 2 of 68

iii

Table of Contents

Acknowledgements .........................................................................................................................1

Executive SummaryEnglish......................................................................................................................................3Tetun Translation ......................................................................................................................7Portuguese Translation ...........................................................................................................12

1 Introduction ................................................................................................................................171.1 Coastal Habitat Mapping ..................................................................................................171.2 Approaches to Coastal Habitat Mapping ..........................................................................181.3 Project Aims & Objectives ................................................................................................19

2 Physical Setting ..........................................................................................................................19

3 Methods ......................................................................................................................................223.1 Input Data .........................................................................................................................223.2 Image Pre-processing ......................................................................................................233.3 Image Classification .........................................................................................................25

3.3.1 Pixel Based ..........................................................................................................253.3.2 Object Based Image Analysis ..............................................................................263.3.3 Manual Interpretation...........................................................................................293.3.4 Fine-scale Mangrove mapping (1:10,000) ............................................................30

3.4 Coastal-Marine Field Surveys ...........................................................................................303.4.1 Coastal Habitats ..................................................................................................303.4.2 Marine Habitats ...................................................................................................313.4.3 Classification Refinement, Validation & Output ....................................................35

4 Results ........................................................................................................................................364.1 Coastal Habitats ...............................................................................................................40

4.1.1 Overview of Coastal Habitats ..............................................................................404.1.2 Beaches & Saltflats ..............................................................................................404.1.3 Coastal Forests (non-mangrove) ..........................................................................414.1.4 Mangroves ...........................................................................................................424.1.5 Open Forest, Woodland & Shrubland ..................................................................444.1.6 Coastal Mapping Accuracy ..................................................................................45

4.2 Marine Habitats ................................................................................................................47

5 Discussion ..................................................................................................................................515.1 Coastal-Marine Mapping ..................................................................................................515.2 Mangroves & Coastal Wetlands ........................................................................................52

5.2.1 Distribution ..........................................................................................................525.2.2 Species Diversity .................................................................................................535.2.3 Community Structure ...........................................................................................555.2.4 Current Uses and Threats ....................................................................................56

5.3 Marine Habitats ................................................................................................................57

iv

6 Capacity-Building & Engagement .............................................................................................586.1 International ‘Reef Check’ Training ...................................................................................586.2 Coastal & Marine Habitat Surveys ....................................................................................586.3 Marine Taxonomy .............................................................................................................596.4 Regional Engagement ......................................................................................................59

7 Emerging Issues & Priorities .....................................................................................................59

8 References..................................................................................................................................61

AppendicesAppendix 1: Timor Leste inshore marine habitat field survey classification system .........................64

Appendix 2: Timor Leste marine habitat mapping field data survey sheet (and MS Access database) ..............................................................................................................65

Appendix 3: Timor Leste coastal vegetation field survey data collection sheet ...............................66

Appendix 4: Terrestrial plant species recorded in the Timor Leste coastal field surveys .................. 67

Appendix 5: Comparison of Timor Leste field survey coastal vegetation groupings with broad mapping classes ..............................................................................................................................68

Appendix 6: ‘Reef Check’ Training and Surveys in Timor Leste .......................................................69

1

Acknowledgements

This project received significant support from organisations and individuals throughout.

We would particularly like to thank the Government of Timor-Leste, and specifically the

Ministry of Agriculture, Forestry and Fisheries for their long-term support in this project.

Within MAFF, we would particularly like to thank Narciso de Carvalho, Anselmo Lopes

Amaral, Adriano "Dani" Fernandes do Karmo, Joaõ D.M. dos Reis, Jose Monteiro,

Carlos de Jesus and Raimundo Mau for assistance. When in Timor-Leste, Mike

Gallagher’s support, as the NT Government representative in Dili, was invaluable.

Dimity Boggs provided valuable input and review of the draft report.

Finally, a very special thankyou to all staff from the Marine and Biodiversity Group

(NRETAS, NT Government) and CDU Tropical Spatial Science Group that provided

important support, input and valuable feedback throughout the project.

Page 3 of 68

334

Project Title: Marine & Coastal Habitat Mapping in Timor Leste Project Leaders: Guy Boggs (CDU), Karen Edyvane (NRETAS, NT Government) Project Participants: Narciso de Carvalho (MAF-Fisheries), Shane Penny (NRETAS), Juno Rouwenhorst (CDU), Peter Brocklehurst (NRETAS), Ian Cowie (NRETAS), Celestino Barreto de Cunha (MAF-Fisheries), Anselmo Lopes Amaral (MAF-Fisheries), Neil Smit (NRETAS), Jose Monteiro (MAF-Fisheries), Pedro Pinto (MAF-Protected Areas), Raimundo Mau (ALGIS), Joao Xavier Amaral (MAF-Fisheries), Lucas Fernandes (MAF-Fisheries).

Introduction The maritime estate of Timor-Leste, encompasses over 700 km of coastline (and the islands of Atauro and Jaco), within a region of globally-significant tropical marine biodiversity. Timor Leste however, also faces major human development challenges – with some of the highest levels of population growth, infant mortality, malnutrition, unemployment, illiteracy and food insecurity in the South East Asia. With many coastal communities highly dependent on coastal resources - knowledge of the distribution, nature and ecological condition of coastal and marine biodiversity and resources is essential to underpin conservation, ecosystem-based planning and sustainable use and also, regional economic development (particularly fisheries, aquaculture and coastal-marine ecotourism). Habitat mapping is vital for the both, the planning of individual Marine Protected Areas (MPAs) and MPA networks, and also, to assist with multiple-use zoning and assisting sustainable coastal resource use and development. While the north coast of Timor-Leste is highly valued for its contribution to local livelihoods, particularly through ecotourism and fisheries-based activities, no broad-scale, habitat mapping currently exists for this region. The primary aim of the present study was to undertake broadscale survey and mapping of the marine and coastal habitats of the north coast of Timor-Leste, to assist with future coastal-marine conservation, planning, and regional sustainable development. Remote Sensing Mapping of the coastal and shallow marine habitats surrounding Timor Leste was undertaken using Landsat TM/ETM+ imagery – using both, object-based and pixel-based analysis techniques. The cloud, tidal and seasonal characteristics of Landsat imagery captured between 2003 and 2006 were identified with two images selected based on low cloud cover, optimal water quality and low tide conditions. These images were analysed using object based image analysis (OBIA) in Definiens® software, an approach that initially segments the imagery and then classifies objects based on their spatial, spectral and textural characteristics. A segmentation approach and class hierarchy developed based on subset study regions was applied to the entire northern coastline and the output mapping edited in ESRI ArcGIS® to create a final map product. Field Surveys Field surveys in November 2007 collected detailed information about the structure and composition of the Timor-Leste northern coastlines marine and coastal habitats. The data from almost 1000 geo-referenced points were analysed using standard geomorphological, benthic marine and terrestrial vegetation classification schemes,

EXECUTIVE SUMMARY

445

based on geomorphic, structural and compositional characteristics (ie. Mather & Bennett 1994, English et al.1997, National Vegetation Information System of Australia). Mangrove communities were also mapped at a fine-scale through interpretation of 1:10,000 aerial photography. Field ground-truthing (of remotely sensed maps) in November 2007, focussed around 6 locations along the Timor Leste coast – Metinaro, Manatuto, Baucau, Com, Jaco and Lore. Detailed surveys of habitats undertaken in the area of proposed Nino Conis Sanatana Marine Park (Com-Tutuala-Lore). The coastal survey collected >1000 points across 5 of the training areas, with existing data collected by the NT Herbarium utilised for Lore. Data collected included locational information, vegetation structure and composition of dominant species (in upper, mid and ground levels), ground cover and simple soil characteristics. Mangrove and saltmarsh specialists (from NRETAS) also undertook detailed floristic surveys within the 6 training areas. The complementary marine survey sampled a total of 724 sites within the 6 training areas, using local boats (and accommodation) from coastal fishing communities. Field data was collected using both geomorphological and biological criteria (adapted from Mumby & Harborne 1999), and undertaken via both, visual census and remote underwater video. Reef condition (ie. bleaching, anchor damage, Crown of Thorns, eutrophication, etc.) was also assessed. Sampling within training areas focussed (where possible) on systematic, cross-shelf transects (perpendicular to the coast) to enable reef profiles to be developed. Marine and Coastal Habitat Map Based on the results of remote sensing analyses and the coastal-marine “field ground-truthing”, a broad scale (1:100k) map of the coastal and marine habitats of the north and east coast of Timor-Leste has been produced. The smallest mapped element is 0.1ha and largest is 1040 ha with a mean polygon size of 10ha (excluding the ‘Deep Water’ class). As expected, the dataset is highly linear, having a maximum ‘width’ of approximately 3km (covering marine and coastal habitats), with much of the mapping being less than 1km in width. The classification identified a total of 6 marine habitat types and 7 coastal habitat types. The marine nearshore zone is characterised by a narrow reef flat (often < 60m wide, but up to almost 1km), dominated by seagrass in shallower water (approximately 2,200ha) and corals in deeper water and on the escarpment (approximately 2,000ha). A mixed-cover class, which included low covers of coral or seagrass and bare areas, occupied 1,250 ha. Within the coastal habitats, mangroves occupied 750ha, the largest stand being in the Metinaro region. Detailed mangrove mapping (1:10,000) identified a total of 20 mangrove communities with clear zonation patterns. Bare areas (including salt flats and beaches) occupied 1,100ha, while the vegetation was dominated by savannas, with 3,750ha of open woodland or shrubland, 1,935ha of open forest or woodland and 1,050ha of coastal forest. Intensive agriculture (2,200ha) and built up areas (1730ha) were also a prominent feature of the northern coastline of Timor Leste. The very limited extent of coral reef, seagrass and mangrove habitats on the north coast of Timor Leste, impose strong limits on available marine resources and levels of harvest (particularly reef fisheries, mangroves) and in the light of increasing human resource use, underscore the urgent need for precautionary and effective conservation management.

556

Importantly, the coastal mapping has revealed significant and ongoing coastal habitat loss in Timor Leste. As such, total mangrove extent has reduced from 9,000ha in 1940, to 3,035ha in 2000 (FA0 2003) to just 1,802 ha recorded in 2008, ie. ~40% loss between 2000-2008, or disturbingly, an approximate 80% loss of total mangrove habitat since 1940. Mangrove trees are harvested for timber and fuel wood and in some instances hinterland mangroves have been removed for brackish water shrimp and/or fish ponds. Validation of the final mapping identified overall accuracies of 83% and 92% for the coastal and marine habitats respectively. However, the broad definition of some habitat classes indicate that these accuracies should be interpreted with some caution and greater value can be found in the comparison of the detailed field survey classes identified within each broader habitat class. Capacity-Building and TL Engagement The project was well-supported by the project partners, the Ministry of Agriculture and Fisheries (MAF) – with a wide range of training outcomes. Training in marine species identification and also, coral reef monitoring was considered a major training priority for MAF staff. Training in survey, marine species identification and ongoing monitoring of coral reef habitats (including fish, invertebrates) was undertaken using the international, ‘ReefCheck’ protocol (www.reefcheck.org). Under the TL ‘ReefCheck’ Project, 3 trainers were trained to deliver ‘ReefCheck’ training and a total of 12 MAF staff were trained – with 3 certified as Eco-Divers. ‘ReefCheck’ training was undertaken using both, the SCUBA-based protocol and the snorkel-based protocol – and with the training delivered both, in English and in Indonesian – using Indonesian training material and products. Two MAF staff have also received training in marine taxonomy, as part of a 5-day, Marine Polychaete Taxonomy Workshop (1-5 October 2007) held in Darwin, by the NT Museum & Art Gallery (NRETAS). The field surveys were also strongly supported by MAF staff (Fisheries, Protected Areas) and ALGIS, with a total of 4 staff assisting the coastal survey and 9 in the marine field survey. Knowledge was shared on terrestrial plant identification and vegetation structural description, as well as GPS data collection techniques and also through a ½ day marine workshop on the use of GPS, depth sounders and remote underwater videos, and the benthic habitat classification. MAF staff also assisted with the organisation of the field logistics (ie. liaison with local communities, boat hire, accommodation, etc.). MAF Priorities & Recommendations 1. In identifying GoTL priorities for further coastal-marine habitat mapping, MAF have

highlighted the need to address priorities for MPA network planning in Timor Leste - particularly, within the context of the broader, MPA network planning in the Lesser Sunda region being undertaken collaboratively with Indonesia (and The Nature Conservancy). To this end, MAF have identified the following priorities for further mapping: (1) west of Dili, including proposed sites for MPAs (high priority); (2) Oecussi (low priority); (3) south coast (low priority). Coastal (terrestrial) mapping and marine mapping were both identified as of equal importance.

667

2. In identifying GoTL mapping priorities, finer-scale mapping for the Nino Konis Santana Marine Park was identified by MAF as the highest habitat mapping priority, building on mapping being undertaken by ALGIS. Areas need to be prioritised for finer-scale mapping and a common classification scheme developed for the Timor Leste region. Capacity building was identified as high priority as it would deliver skills and knowledge for MAF (Fisheries/ALGIS) staff to specifically undertake mapping and spatial data analysis at both, the fine and broad-scale.

3. MAF have identified capacity-building (training and equipment) in coastal-marine

habitat mapping between Timor Leste, Australia (Northern Territory) and Indonesia (Nusa Tengarra Timor), as a high priority. Further, the need for a common classification scheme (and scales of information) and methods, across the region has also been identified as a key priority. This could be implemented through a regional, coastal mapping workshop, with participants from MAF (Fisheries/ALGIS), Nusa Tengarra Timor and the Northern Territory. This workshop would define terms of reference between the 3 countries on classification schemes and priorities and also, develop a plan of action. MAF also identified the need to train 2-3 staff in marine and coastal GIS and remote sensing techniques, in partnership with CDU.

4. Database training and knowledge management (and also, awareness of Intellectual

Property) is a priority area for training of MAF (and other GoTL agency staff). With knowledge capture and the building of GIS and knowledge systems, training is required by MAF staff in database management, data standards, Intellectual Property and formal data agreements. MAF have also identified database training and knowledge management as a priority area for capacity-building and training of MAF and GoTL staff.

5. MAF recognise the importance of Intellectual Property and knowledge management

and recommend that all requests for data/information arising from work currently undertaken by MAF (and the Government of Timor Leste) (including this partnership program) should be subject to a formal data agreement.

6. Under the CTI National Plan of Action for Timor Leste, the Wetar Strait is priority

‘seascape’ requiring deep water habitat mapping. MAF, however, do not consider this a mapping priority, given the current priorities and lack of capacity and resources within MAF.

7. Given the success of the ‘ReefCheck’ program in Timor Leste, MAF have identified

the need to support and expand the existing MAF ‘ReefCheck’ survey program and team in Timor Leste. This includes establishing regular monitoring, 2-4 times year per site and increasing the number of sites across the northern coast. It was recommended that: (a) the reef survey program use ‘ReefCheck’ and also, other survey methods; (b) include impacted sites; and (c) establish regular monitoring sites within the Nino Konis Santana Marine Park (both, inside and outside the park).

8. MAF have identified the need to build on the TL ‘ReefCheck’ program (and results)

and develop a school-based, marine education and public awareness program, to highlight the status, condition and threats facing coral reefs in Timor Leste. In the Com-Tutuala region, this education campaign could be a component of a broader, Marine Park community outreach and education program.

778

Projetu Naran: Mapeamentu ba Habitat iha tasi-Ibun no Tasi-Laran Timor-Leste Lider sira Projetu nian: Guy Boggs (CDU), Karen Edyvane (NRETAS) Partisipante sira iha Projetu ne’e: Narciso de Carvalho (MAF-Fisheries), Shane Penny (NRETAS), Juno Rouwenhorst (CDU), Peter Brocklehurst (NRETAS), Ian Cowie (NRETAS), Celestino Barreto de Cunha (MAF-Fisheries), Anselmo Lopes Amaral (MAF-Fisheries), Neil Smit (NRETAS), Jose Monteiro (MAF-Fisheries), Pedro Pinto (MAF-Protected Areas), Raimundo Mau (ALGIS), Joao Xavier Amaral (MAF-Peska), Lucas Fernandes (MAF-Peska). Introdusaun Propriedade marítimu Timor-Leste nian engloba área tasi-ibun (ka kosteira) liu kilómetru 700 (no tan illa Atauro no Jaco), iha rejiaun ida ne’ebé globalmente iha signifikadu bio-diversidade marítima tropikál. Maibé, Timor-Leste hasoru mós dezafiu boot ida kona-ba dezenvolvimentu umanu – ho kreximentu populasionál balu ne’ebé aas tebes iha Sudeste Asia nia laran, mortalidade infantil, malnutrisaun, dezempregu, analfabetizmu, falta seguransa ai-han nian. Ho populasaun barak iha zona kosteira ne’ebé depende ba rekursus iha tasi-ibun – importante tebes hasa’e koñesimentu kona-ba distribuisaun, bio-diversidade tasi-ibun no tasi-laran (kosteira no marina) nia kondisaun natural no ekolójika atu bele sustenta protesaun, planeamentu ida ne’ebé bazeia ba eko-sistema no uzu sustentavel no mós dezenvolvimentu ekonómiku rejionál (liliu peska, aqua-kultura no eko-turizmu kosteiru no mariñu). Mapeamentu kona-ba habitat ne’e importante tebes hodi halo planeamentu ba Area Mariña Protejida idaidak (MPAs) no ba rede MPA ninian, no mós atu fó asisténsia ba zoneamentu uzu-múltiplu no asisténsia ba uzu sustentavel no dezenvolvimentu ba rekursus sira iha tasi-ibun. Kosta norte Timor-Leste nian iha valór boot tebes tanba kontribuisaun ne’ebé nia fó ba populasaun lokál nia moris, liliu husi eko-turizmu no atividade sira ne’ebé bazeia ba peska, maibé sei dauk iha mapeamentu no peskiza ida iha eskala luan ba rejiaun ne’e. Estudu ida-ne’e nia objetivu prinsipál mak atu halo mapeamentu no peskiza ba habitat sira iha tasi-ibun no tasi-laran Timor-Leste nian hodi fó asisténsia ba konservasaun, planeamentu no dezenvolvimentu sustentavel iha rejiaun ne’e iha futuru. Sensoriamentu husi distánsia Uza Land sat TM/ETM+ imajen – objetu baze no téknika análize ne’ebé bazeia ba pixel – halo ona mapeamentu ba habitat mariñu sira iha tasi-ibun hadulas Timor-Leste. Identifika ona kaloan, laloran no karaterístika temporál husi Landsat nia imajen ne’ebé hasai entre 2003 no 2006 no hili ona imajen rua (2), bazeia ba kaloan menus, bee nia kualidade di’ak tebes no kondisaun laloran nian badak. Uza objetu análize imajen ninian (object based image analysis, OBIA), husi Definiens® software, halo ona análize ba imajen hirak ne’e. Prosesu ne’e uluk nanain fahe imajen sira iha grupu hafoin tuir mai klasifika objetu hirak ne’e tuir sira-nia karaterístika espasiál, espectral (spectral) no textural (textural). Aplika ona prosesu segmentasaun (fahe ba grupu) no klase kategoria nian, ho baze iha estudu subkonjuntu iha rejiaun sira, ba kosta norte tomak Timor-Leste nian, no halo sai mapeamentu ida ne’ebé editadu iha ESRI ArcGIS ® atu kria mapa finál ida. Peskiza sira iha terrenu

889

Peskiza sira ne’ebé hala’o iha terrenu, iha Novembru 2007, halibur ona informasaun ho detalle kona-ba estrutura no kompozisaun liña kosteira iha parte norte Timor-Leste nian no sira-nia habitat. Uza padraun geomorfológico, bêntico marinho no sistema klasifikasaun ba ai-moris iha rai leten, bazeia ba geomórfico, estrutura no kompozisaun sira-nia karaterístika (ezemplu, Mather & Bennett 1994, English et al.1997, National Vegetation Information System of Australia), halo ona análize ba dadus husi pontos georreferenciados besik rehun ida (1000). Halo mós mapeamentu ba komunidade ai-tasi iha fine-scale ida husi interpretasaun ba fotografia aérea 1:10,000. Ground-trutting ne’ebé hala’o iha Novembru 2007 (mapas de teledetecção), foka ba fatin neen (6) nune’e iha kosta Timor-Leste nian - Metinaro, Manatuto, Baucau, Com, Jaco no Lore. Halo ona peskiza detalladu ba habitat sira iha área propostu Parke Mariñu Nino Konis Santana (Com-Tutuala-Lore). Peskiza ne’ebé halo iha tasi-ibun ne’e halibur pontu rehun ida (1000) husi área formasaun lima, ho dadus ne’ebé NT Herbarium halibur hodi uza ba Loré. Dadus detalladu sira inklui informasaun kona-ba fatin, estrutura vegetação nian no espésie dominante sira-nia kompozisaun (iha rai tetuk, médiu no aas), no karaterístika kobertura rai nian no rai simples. Espesialista sira kona-ba ai-oan moris iha tasi-ibun no marisma (salt marsh), husi NRETAS, halo mós levantamentu florístico (floristic) iha área formasaun neen (6) nia laran. Uza roho (no fatin-hela) husi populasaun peskadór lokál, peskiza mariñu komplementár halibur fatin hamutuk 724 iha área neen (6) formasaun ninian. Uza kritériu geomorfológico no biolójiku (adaptadu husi MUMBY & Harborne 1999), sira halibur dadus iha terrenu, no uza censo visual no vídeo subaquático husi distánsia. Halo mós avaliasaun ba kondisaun au-ruin ninian (branqueamento, estragu husi âncora, Crown of Thorns, eutrophication no seluk tan). Ezemplar ka amostra sira iha área formasaun nia laran (se bele) foka ba transecções sistemáticas, cross-prateleira (liña perpendikular tasi-ibun ninian) atu bele dezenvolve au-ruin sira-nia perfil. Mapa habitat Kosteiru no Mariñu Bazeia ba rezultadu sira husi análize sensoriál husi distánsia no bazeia ba “field ground-truthing” kosteiru-mariñu, halo ona mapa ida ho eskala boot tebes (1:100 k) kona-ba habitat kosteiru no mariñu iha kosta norte no leste Timor-Leste nian. Elementu mapeadu ki’ik liu mak 0.1ha no boot liu mak 1040 ha, ho nia luan média polígono 10ha (ne’e la sura ho kategoria ‘Deep Water' ka bee-kle’an). Hanesan ita hein tiha ona, dadus hirak ne’e lineár tebetebes, ho nia ‘largura’ másimu 3 km (abranje habitat tasi-ibun no tasi-laran nian), ho mapeamentu barak liu mak la liu kilómetru 1 largura (1 km). Klasifikasaun ne’e hetan habitat mariñu oin neen (6) no habitat kosteiru oin hitu (7). Mariñu iha zona tasi-ninin nian halo parte au-ruin kloot ida no tetuk/belar (ne’ebé jeralmente iha metru 60 largura, maibé to’o besik kilometru 1), ne’ebé duut-tasi mak domina iha bee tetuk ka planu (besik 2.200 ha) no au-ruin iha bee kle’an liu no iha escarpa (escarpment) nia leten (besik 2.000 ha). Mistura klase kobertura, ne’ebé inklui menus kobertura ba au-ruin ka duut-tasi no área sira mamuk, okupa 1.250 ha. Iha habitat kosteiru sira-nia laran, ai-tasi sira okupa 750ha. Detalle husi mapeamentu kona-ba ai-tasi sira (1:10.000) identifika komunidade ai-tasi ruanulu (20) ho padraun klaru kona-ba zoneamentu. Area mamuk sira (inklui kopartimentu masin nian no praias) okupa 1.100 ha, enkuantu savana sira (ai-laran) ne’ebé domina vegetação, ho 3.750ha ai-laran ka de floresta aberta ka arbustu, 1.935 ha floresta aberta ka floresta no 1.050 ha floresta kosteira. Area ne’ebé iha agrikultura intensiva (2.200 ha) no área ne’ebé populasaun hela (1730h), halo mós parte hanesan karaterístika importante tebes kosta norte Timor-Leste nian.

9910

Habitat au-ruin, duut-tasi no ai-tasi nia estensaun limitadu iha kosta norte Timor-Leste nian inpoin limitasaun boot ba rekursus ne’ebe eziste iha tasi-laran no nivel colheita nian (harvest) (liliu au-ruin, ai-tasi sira); tanba utilizasaun rekursus umanus nian aumenta, presiza harii lalais jestaun kona-ba konservasaun preventiva no efikás. Importante liu, mapa tasi-ibun nian hatudu katak habitat iha tasi-ibun lakon barak ona no kontinua lakon nafatin iha Timor-Leste. Nune’e, ai-tasi nia estensaun totál redús husi hectare 9,000 iha 1940 ba hectare 3,035 iha 2000 (FAO 2003) no tun ba hectare 1,802 ne’ebé rejista iha 2008, ezemplu: lakon pursentu 40 (40%) entre 2000-2008, ka aat liu tan ne’e, lakon ai-tasi nia habitat totál besik 80% hahú husi 1940. Ema tesi ai-tasi atu halo ai-kabelak no ai-maran, no dala ruma hasai tiha ai-tasi iha rai leten hodi hakiak boek-oan iha bee merak no/ka halo debun ba ikan. Validasaun husi mapeamentu final identifika katak iha jerál hetan 83% ezatidaun (accuracies) ba habitat kosteira no 92% ezatidaun ba habitat mariña. Maibé, definisaun jerál ba habitat nia kategoria balu hatudu katak ita presiza interpreta ezatidaun hirak ne’e ho kuidadu, no bele hetan valor boot liu bainhira halo komparasaun entre levanatamentu husi kategoria peskiza iha terrenu iha klase habitat idaidak ne’ebé luan liu. Kapasitasaun no TL nia Kometimentu Projetu ne’e simu apoiu di’ak husi ninia parseiru sira, Ministériu Agrikultura no Peska (MAF) – ho rezultadu lubun ida husi formasaun. Treinu kona-ba oinsá atu identifika espésie mariña no mós monitorizasaun ba au-ruin, ne’e konsidera hanesan prioridade boot ida ba funsionáriu MAF sira-nia treinamentu. Uza protokolu internasionál ‘ReefCheck’ (www.reefcheck.org), halo ona treinu kona-ba peskiza, identifikasaun ba espésie mariña sira no monitorizasaun permanente ba au-ruin nia habitat sira (inklui ikan, invertebrados). Iha Projetu TL 'ReefCheck' nia laran, treinadór na’in tolu simu ona treinu atu hala’o treinu kona-ba "ReefCheck’, no funsionáriu MAF hamutuk ema na’in 12 mak simu ona treinu ne’e – ne’ebé na’in tolu hetan sertifikadu hanesan Eco-Divers. Uza protokolu ne’ebé bazeia ba SCUBA no protokolu ne’ebé bazeia ba snorkel, hala’o ona treinu kona-ba ‘ReefCheck’ – treinu ne’e hala’o iha lia-Indonesia no lia-Inglés – no uza material no produtu sira husi Indonesia. MAF nia funsionáriu na’in rua simu mós treinu kona-ba taxonomia marinha, hanesan parte iha workshop ida ne’ebé Museum & Art Gallery (NRETAS) hala’o iha Darwin, NT, durante loron lima (1-5 Outubru 2007) kona-ba Marine Polychaete Taxonomy. Levantamentu iha terrenu mós simu apoiu maka’as husi MAF nia funsionáriu (Peska, Area Protejida) no ALGIS, ho funsionáriu hamutuk na’in haat (4) ne’ebé fó asisténsia ba peskiza kosteira no na’in sia (9) ba peskiza iha área mariña nian. Sira fahe koñesimentu identifikasaun ninian kona-ba ai-oan moris sira iha rai leten (plantas terrestres) no deskrisaun estruturál kona-ba ai-laran (vegetação), no mós téknika kona-ba halibur dadus GPS no mós liu husi workshop loron-balu nia laran kona-ba oinsá atu uza GPS, profundidade acústica (depth sounders) video submarinu no klasifikasaun ba habitats bênticos (benthic habitat). MAF nia funsionáriu sira mós simu apoiu kona-ba organizasaun lojístika iha terrenu (ezemplu, ligasaun ho komunidade lokál sira, aluga roho, toba-fatin no seluk tan). MAF nia Prioridade no Lia-Menon

101011

1. Bainhira identifika GoTL nia prioridade sira atu halo tan mapa ba riku-soin iha tasi-laran no tasi-ibun sira-nia habitat, MAF hateten katak iha nesesidade boot atu fó prioridade ba MPA nia rede planeamentu iha Timor-Leste laran – liliu, iha kontestu jerál, tanba MPA nia rede planeamentu iha rejiaun Lesser Sunda hala’o daudaun ona ho Indonesia (no ho Konservasaun Natureza). Kona-ba ne’e, MAF identifika ona prioridade sira tuir mai ne’e, ne’ebé atu halo tan mapa: (1) parte loromunu Dili nian, inklui área hirak ne’ebé propoin ona ba MPA (prioridade boot); (2) Oecusse (prioridade ki’ik); (3) kosta tasi-mane nian (prioridade ki’ik). Mapa kosta nian (rai leten) no mapa tasi-laran nian identifika ona katak iha importánsia hanesan.

2. Bainhira identifika GoTL nia prioridade sira, MAF identifika mapeamentu eskala finu

ba Parke Mariñu Nino Konis Santana hanesan mapeamentu habitat nian ne’ebé iha prioridade boot liu. ALGIS harii daudaun ona baze ba mapeamentu ne’e. Presiza fó prioridade ba área sira ne’ebé atu halo mapeamentu eskala finu no dezenvolve sistema klasifikasaun komún ba rejiaun Timor-Leste nian. Identifika ona katak prioridade boot mak kapasitasaun hodi hato’o kompeténsia no koñesimentu ba MAF nia funsionáriu sira (Peska/ALGIS) atu halo espesifikamente mapa no análize ba dadus espasiál sira, iha eskala finu no mós eskala globál.

3. MAF identifika ona kapasitasaun (treinu no ekipamentu) kona-ba mapeamentu ba habitat sira tasi-laran no tasi-ibun entre Timor-Leste, Australia (Territoriu Norte) no Indonesia (Nusa Tengara Timur) hanesan prioridade ida boot. Alende ne’e, identifika mós nesesidade atu estabelese sistema no métodu identifikasaun (no eskala informasaun nian) komún ba rejiaun tomak hanesan prioridade prinsipál ida. Bele implementa ne’e husi workshop ida kona-ba mapeamentu ba tasi-laran no tasi-ibun, rejionál, ho partisipante sira husi MAF (Peska/ALGIS), Nusa Tengara Timur no Territóriu Norte. Workshop ne’e bele define termu referénsia kona-ba sistema klasifikasaun no prioridade sira entre nasaun tolu ne’e, no mós dezenvolve planu asaun ida. MAF identifika mós katak iha nesesidade atu fó treinu ba funsionáriu na’in 2-3 kona-ba GIS tasi-laran no tasi-ibun no téknika sensoriamentu remotu, husi parseria ho CDU.

4. Treinu kona-ba banku-dadus no jestaun koñesimentu nian (no mós koñesimentu

kona-ba Propriedade Intelektual) ne’e área prioridade ida ne’ebé atu fó treinu ba MAF (no ba funsionáriu sira husi GoTL nia ajénsia sira seluk). Ho koñesimentu ne’ebé hetan no konstrusaun sistema GIS no koñesimentu nian, MAF nia funsionáriu sira presiza simu treinu kona-ba jestaun banku-dadus, padraun dadus nian, Propriedade Intelektual no akordu formál kona-ba dadus. MAF identifika mós treinu kona-ba banku-dadus no jestaun koñesimentu nian hanesan área prioridade kapasitasaun no treinu ninian ba MAF no GoTL nia funsionáriu sira.

5. MAF rekoeñse katak Propriedade Intelektual no jestaun koñesimentu nian importante no hameno katak pedidu hotu-hotu ne’ebé mai husi servisu ne’ebé hala’o daudaun husi MAF (no Governu Timor-Leste) (inklui parseria husi programa ne’e) tenke submete ba akordu formál ida kona-ba dadus.

6. Tuir Planu Asaun Nasionál CTI ba Timor-Leste, Wetar Strait (Estreitu Wetar) ne’e ‘seascape’ prioritáriu ida ne’ebé presiza mapeamentu ba habitat iha bee-kle’an. Maibé, MAF la konsidera ne’e hanesan prioridade atu halo mapa tanba prioridade sira ne’ebé iha no falta kapasidade no rekursus iha MAF nia laran.

111112

7. Haree ba susesu programa ‘ReefCheck’ nian iha Timor-Leste, MAF identifika katak iha nesesidade atu fó apoiu no haluan MAF nia programa no ekipa peskiza ‘ReefCheck’ ne’ebé eziste iha Timor-Leste. Ida-ne’e inklui monitorizasaun regulár, dala 2-4 tinan ida ba fatin ida, no hasa’e númeru fatin nian iha kosta tasi-feto. Hameno katak: (a) atu uza programa ‘ReefCheck’ no mós métodu sira seluk hodi halo peskiza ba au-ruin; (b) atu inkui fatin hirak ne’ebé hetan impaktu; no (c) atu define fatin atu halo monitorizasaun regulár iha Parke Mariñu Nino Konis Santana (iha parke nia laran no nia li’ur).

8. MAF identifika katak iha nesesidade atu dezenvolve iha programa (no rezultadu

sira) ‘ReefCheck’ Timor-Leste ne’ebé mak iha ona no halo programa kona-ba sensibilizasaun públika no edukasaun kona-ba riku-soin tasi nian iha eskola sira hodi hatudu au-ruin sira iha Timor-Leste nia estatutu, kondisaun no ameasa ne’ebé sira hasoru. Iha rejiaun Com-Tutuala, kampaña edukasaun ne’e bele sai hanesan komponente ida husi programa edukasaun komunitária kona-ba Parke Mariñu.

121213

Título do Projecto: Mapeamento de Habitats Marinhos e Costeiros em Timor-Leste Líderes do Projecto: Guy Boggs (CDU), Karen Edyvane (NRETAS, Governo do

Northern Territory) Participantes no Projecto: Narciso Almeida de Carvalho (MAP – Pescas), Shane Penny (NRETAS), Juno Rouwenhorst (CDU), Peter Brocklehurst (NRETAS), Ian Cowie (NRETAS), Celestino da Cunha Barreto (MAP – Pescas), Anselmo Lopes Amaral (MAP – Pescas), Neil Smit (NRETAS), José Monteiro (MAP – Pescas), Pedro Pinto (MAP – Áreas Protegidas), Raimundo Mau (MAP – ALGIS), João Xavier Amaral (MAP – Pescas), Lucas Fernandes (MAP – Pescas). Introdução Timor-Leste tem uma linha de costa de mais de 700 km (e as ilhas de Ataúro e Jaco), numa região de biodiversidade marinha tropical globalmente significativa. No entanto, Timor-Leste depara-se com enormes desafios de desenvolvimento humano, com os maiores níveis de crescimento populacional, mortalidade infantil, malnutrição, desemprego, iliteracia e insegurança alimentar do Sudeste Asiático. Com muitas comunidades costeiras muito dependentes dos recursos costeiros, o conhecimento da distribuição, natureza e condição ecológica da biodiversidade e dos recursos costeiros e marinhos é essencial para sustentar a conservação, planeamento com base no ecossistema e o uso sustentável e também o desenvolvimento económico regional (particularmente as pescas, aquacultura e o ecoturismo marinho e costeiro). O mapeamento de habitats é essencial para o planeamento de Áreas Marinhas Protegidas (AMP) individuais e também para o planeamento das redes de AMP’s, e ainda no apoio à definição de zonas de usos múltiplos e ao uso e desenvolvimento sustentável dos recursos costeiros. Apesar da costa Norte de Timor-Leste ser altamente valorizada pela sua contribuição aos meios de vida locais, em particular através do ecoturismo e actividades piscatórias, não existe nenhum mapa em larga escala de habitats para esta região. O principal objectivo deste estudo é realizar um levantamento e mapeamento de larga escala para os habitats marinhos e costeiros na costa Norte de Timor-Leste, para apoiar os esforços futuros de conservação costeira e marinha, de planeamento e de desenvolvimento regional sustentável. Detecção Remota O mapeamento dos habitats costeiros e marinhos de baixa profundidade junto a Timor-Leste foi realizado recorrendo a imagens Landsat TM/ETM+, usando técnicas de análise com base em objectos e em pixéis. Identificaram-se as condições de nebulosidade, das marés e características sazonais das imagens Landsat obtidas entre 2003 e 2006, sendo seleccionadas duas imagens com condições de pouca nebulosidade, óptima qualidade da água e baixa-mar. Estas imagens foram analisadas recorrendo a uma análise de imagens com base em objectos (OBIA – object based image analysis) através do software Definiens®, uma abordagem que inicialmente segmenta a imagem e depois classifica os objectos com base nas suas características espaciais, espectrais e de textura. Uma abordagem de segmentação e hierarquia de classes desenvolvida com base em sub-regiões de estudo foi aplicada a toda a linha de costa Norte e o mapa resultante foi editado em ESRI ArcGIS® para criar um mapa final. Levantamentos no campo

131314

Em Novembro de 2007, durante as campanhas de campo, foi colectada informação detalhada acerca da estrutura e composição dos habitats marinhos e costeiros na linha de costa Norte de Timor-Leste. Os dados de quase 1000 pontos georreferenciados foram analisados usando esquemas padrão de classificação geomorfológica e de vegetação marinha bentónica e terrestre, com base nas características geomórficas, estruturais e de composição (ie. Mather & Bennet 1994, English et al. 1997, National Vegetation Information System of Australia). As comunidades de mangal foram também mapeadas a fina escala, através da interpretação de fotografias aéreas à escala 1:10,000. A verificação no terreno (de mapas de detecção remota) em Novembro de 2007, concentrou-se ao redor de 6 locais ao longo da costa Norte de Timor-Leste – Metinaro, Manauto, Baucau, Com, Jaco e Lore. Foram feitos levantamentos detalhados na zona do proposto Parque Marinho Nino Konis Santana (Com-Tutuala-Lore). O levantamento costeiro obteve mais de 1000 pontos em 5 das áreas de treino, com informação existente recolhida pelo Herbário do Northern Territory (NT Herbarium) usada para Lore. Os dados recolhidos incluíam informações de localização, estrutura vegetal e composição de espécies dominantes (em níveis superiores, médios e inferiores), cobertura vegetal e características simples do solo. Especialistas em mangais (do NRETAS) também efectuaram levantamentos detalhados da flora nas 6 áreas de treino. O levantamento marinho complementar amostrou um total de 724 locais nas 6 áreas de treino, recorrendo a embarcações (e alojamento) locais, das comunidades piscatórias costeiras. Os dados de campo foram recolhidos usando critérios geomorfológicos e biológicos (adaptados de Mumby & Harborne 1999), recorrendo a censos visuais e vídeo subaquático remoto. As condições dos recifes (ie. branqueamento, danos de ancoragem, eutrofização, coroas de espinhos, etc.) foram também avaliadas. A amostragem nas áreas de treino concentrou-se (quando possível) em transectos transversais (perpendiculares à linha de costa) sistemáticos para permitir o desenvolvimento de perfis dos recifes. Mapa de Habitats Marinhos e Costeiros Com base nos resultados da análise dos dados obtidos remotamente e da verificação de campo foi produzido um mapa de larga escala (1:100,000) dos habitats costeiros e marinhos das costas Norte e Este de Timor-Leste. O elemento mapeado mais pequeno tem 0.1ha e o maior 1040ha sendo que a média do tamanho dos polígonos é de 10 ha (exceptuando a classe “águas profundas”). Como esperado, o conjunto de dados é altamente linear, tendo uma “largura” máxima de aproximadamente 3km (cobrindo habitats marinhos e costeiros), sendo que a maior parte do mapa tem menos de 1km de largura. A classificação identificou um total de 6 tipos de habitats marinhos e 7 tipos de habitats costeiros. A zona marinha perto da linha de costa é caracterizada por uma estreita zona plana do recife de coral (reef flat) (normalmente <60m de largura, mas até um máximo de quase 1km), dominada por ervas marinhas em águas menos profundas (aproximadamente 2,200ha) e coral em águas mais profundas e nas escarpas (aproximadamente 2,000ha). A classe de cobertura mista, que inclui baixa cobertura de coral ou de ervas marinhas e áreas descobertas, ocupa 1,250ha. Dentro dos habitats costeiros, os mangais ocupam 750ha, sendo o maior pé na região de Metinaro. O mapeamento detalhado dos mangais (1:10,000) identificou um total de 20 comunidades de mangal com padrões zonais claros. Áreas descobertas (incluindo lagos de sal e praias) ocupavam 1,100ha enquanto a vegetação era dominada por savanas. Com 3,750ha de mata aberta, 1,945ha de floresta aberta e 1,050ha de floresta costeira. A agricultura

141415

intensiva (2,200ha) e as áreas construídas (1,730ha) eram também uma característica proeminente da costa Norte de Timor-Leste. A pequena extensão dos habitats de recifes de coral, ervas marinhas e mangais na costa Norte de Timor-Leste impõem fortes limites aos recursos marinhos disponíveis e aos níveis de exploração (particularmente nas pescas de recife e nos mangais) e à luz de um cada vez maior uso humano dos recursos sublinha a necessidade urgente de uma gestão de conservação efectiva e precautória. Muito importante é a revelação de perda significativa e contínua de habitats em Timor-Leste. A extensão total de mangal diminuiu de 9,000ha em 1940 para 3,035ha em 2000 (FAO 2003) para apenas 1,802ha identificados em 2008, i.e. uma perda de cerca de 40% no período 2000-20008, ou uns alarmantes cerca de 80% desde 1940. As árvores dos mangais são exploradas para madeira de construção e combustível e em alguns casos os mangais mais interiores foram removidos para aquacultura de peixes e camarão. A validação dos mapas finais identificou precisões de 83% e 92% para os habitats costeiros e marinhos, respectivamente. No entanto, a definição abrangente de algumas classes de habitats, indicam que estas precisões devem ser interpretadas com algum cuidado e que é preferível comparar as classes detalhadas dos levantamentos de campo identificadas dentro de cada classe de habitat. Desenvolvimento de Capacidades e Envolvimento de TL O projecto teve muito apoio dos parceiros, do Ministério da Agricultura e Pescas (MAP) – com um largo espectro de resultados de formação. Considerou-se que a formação na identificação de espécies marinhas e na monitorização dos recifes de coral era prioritária para os funcionários do MAP. Foi oferecida formação em levantamentos, identificação de espécies marinhas e monitorização contínua dos habitats de recifes de coral (incluindo peixes e invertebrados), recorrendo ao protocolo internacional “ReefCheck” (www.reefcheck.org). No Projecto ReefCheck de Timor-Leste, formaram-se 3 formadores para dar formação no referido protocolo e um total de 12 funcionários do MAP receberam formação – sendo que 3 obtiveram certificação de Eco-Divers. A formação ReefCheck foi dada usando os protocolos de mergulho com escafandro e de apneia, em duas línguas (Inglês e Bahasa Indonésio, usando os materiais e produtos didácticos indonésios). Dois funcionários do MAP também receberam formação em taxonomia marinha, como parte de um Seminário de Taxonomia de Poliquetas Marinhos de 5 dias (1-5 Outubro 2007) organizado em Darwin pelo NT Museum & Art Gallery (NRETAS). Os levantamentos de campo também foram altamente apoiados pelos funcionários do MAP (Pescas, Áreas Protegidas e ALGIS) com um total de 4 funcionários a apoiarem o levantamento costeiro e 9 no levantamento marinho. Foram transmitidos conhecimentos em identificação de plantas terrestres e descrição estrutural de vegetação, bem como técnicas de recolha de dados GPS e também durante um seminário de ½ dia sobre o uso de GPS, sondas de profundidade e vídeos remotos subaquáticos, e classificação de habitats bênticos. Os funcionários do MAP também deram apoio à organização da logística de campo (i.e. ligação com as comunidades locais, aluguer de embarcações, alojamento, etc.).

151516

Prioridades e Recomendações do MAP 1. Ao identificar as prioridades do Governo de Timor-Leste para mais mapeamento de

habitats costeiros e marinhos, o MAP sublinhou a necessidade de abordar as prioridades de planeamento de AMP’s em Timor-Leste – particularmente no contexto do planeamento mais abrangente da região das Ilhas Menores de Sunda, levada a cabo em colaboração com a Indonésia (e The Nature Conservancy). Neste sentido o MAP identificou as seguintes prioridades para mais mapeamento: (1) Oeste de Dili, incluindo zonas propostas para protecção (alta prioridade); (2) Oecussi (baixa prioridade); (3) Costa Sul (baixa prioridade). Considerou-se que o mapeamento costeiro (terrestre) e o mapeamento marinho tinham igual prioridade.

2. Ao identificar as prioridades de mapeamento do Governo de Timor-Leste, o mapeamento a fina escala do Parque Nacional Nino Konis Santana foi identificado como a prioridade máxima de mapeamento de habitats, complementando o mapeamento que se encontra a ser feito pelo ALGIS. Têm de se identificar as áreas prioritárias para os mapas de fina escala e desenvolver um esquema comum de classificação para a região de Timor-Leste. A capacitação foi também identificada como uma alta prioridade, uma vez que daria capacidades e conhecimentos aos funcionários do MAP (Pescas/ALGIS) para que realizem os mapeamentos e análise espacial de larga e fina escala.

3. O MAP identificou a capacitação (formação e equipamentos) no mapeamento de habitats costeiros e marinhos entre Timor-Leste, Austrália (Northern Territory) e Indonésia (Nusa Tenggara Timur), como sendo de alta prioridade. Para além disto, o desenvolvimento de um esquema de classificação, escalas de informação e métodos comuns à região, também foi identificado como prioritário. Isto poderia ser alcançado com um seminário regional de mapeamento costeiro, com participantes do MAP (Pescas/ALGIS), Nusa Tenggara Timur e Northern Territory. Este seminário definiria entre os 3 países os termos de referência dos esquemas de classificação e prioridades, e também desenvolveria um plano de acção. O MAP também identificou a necessidade de formar 2-3 funcionários em sistemas de informação geográfica (GIS) e técnicas de detecção remota marinha e costeira, em parceria com a CDU.

4. Formação em bases de dados e gestão de conhecimento (e também, conhecimentos de Propriedade Intelectual) é uma prioridade na área da formação dos recursos humanos do MAP (e de outras agências governamentais de Timor-Leste). Com a apreensão de conhecimentos e a construção dos sistemas de GIS e conhecimento, é necessária formação em gestão de bases de dados, padronização de dados, Propriedade Intelectual e acordos formais de dados. O MAP também identificou a formação em bases de dados e gestão de conhecimentos como uma área prioritária na capacitação e formação dos recursos humanos do MAP e do Governo de Timor-Leste.

5. O MAP reconhece a importância da Propriedade Intelectual e a gestão de conhecimentos e recomenda que todos os pedidos de dados/informação decorrentes de trabalhos em curso no MAP (e no Governo de Timor-Leste) (incluindo esta parceria) deveriam estar sujeitos a um acordo formal de dados.

161617

6. Sob o Plano de Acção Nacional para o CTI, o Estreito de Wetar é uma “paisagem marinha” prioritária que requer mapeamento de habitats de águas profundas. No entanto, o MAP não considera este mapeamento como sendo prioritário, devido às prioridades correntes e à falta de capacidade e recursos do MAP.

7. Dado o sucesso do programa “ReefCheck” em Timor-Leste, o MAP identificou a necessidade de apoiar e expandir o programa e equipa existentes. Isto inclui o estabelecimento de monitorização regular, 2-4 vezes por ano em cada local e aumentando o número de locais na costa Norte. Recomendou-se que: (a) o programa de avaliação de recifes use o protocolo “ReefCheck” e também outros meios de avaliação; (b) inclua locais com impactos negativos; (c) se estabeleçam locais de monitorização regular dentro e fora do Parque Nacional Nino Konis Santana.

8. O MAP identificou a necessidade de aumentar o programa (e resultados) “ReefCheck” em Timor-Leste e de desenvolver um programa de educação e sensibilização ambiental, nas escolas, que realcem o estado, condição e ameaças aos recifes de coral de Timor-Leste. Na região Com-Tutuala, esta campanha educativa poderia ser uma componente de um programa de sensibilização e educação comunitária do Parque Nacional.

17

1 INTRODUCTION

1.1 Coastal Habitat Mapping

The maritime estate of Timor-Leste, encompasses over 700 km of coastline (and the

islands of Atauro and Jaco), within a region of globally-significant tropical marine

biodiversity. Timor Leste however, also faces major human development challenges –

with some of the highest levels of population growth, infant mortality, malnutrition,

unemployment, illiteracy and food insecurity in the South East Asia. With many coastal

communities highly dependent on coastal resources - knowledge of the distribution,

nature and ecological condition of coastal and marine biodiversity and resources is

essential to underpin conservation, ecosystem-based planning and sustainable use and

also, regional economic development (particularly fisheries, aquaculture and coastal-

marine ecotourism).

The broad scale mapping of coastal habitats is a fundamental requirement in effectively

managing the coastal environment (Phinn et al. 2005). Habitat maps document the

habitat types present and their statistics and allow the identification of environmentally

sensitive areas and hot spots of habitat diversity (Mumby & Harborne 1999). Habitat

mapping is vital for the both, the planning of individual Marine Protected Areas (MPAs)

and MPA networks, and also, to assist with multiple-use zoning and assisting

sustainable coastal resource use and development (Beech et al. 2008).

While the north coast of Timor-Leste is highly valued for its contribution to local

livelihoods, particularly through ecotourism and fisheries-based activities, no broad-

scale, habitat mapping currently exists for this region. As such, there is limited baseline

information available on the nature, composition, structure, distribution and status of

nearshore coastal and marine habitats in Timor Leste. The primary aim of the present

study was to undertake broadscale survey and mapping of the marine and coastal

habitats of the north coast of Timor-Leste, to assist with future coastal-marine

conservation, planning, and regional sustainable development.

Page 15 of 68

18

1.2 Approaches to Coastal Habitat Mapping

Satellite based remote sensing has been used to map a range of marine and coastal

attributes, from habitat mapping and change detection to measurements of water

quality and productivity in the marine environment (Diaz et al. 2004; Malthus & Mumby

2003; Sanderson 2001). The mapping application will primarily drive the selection of an

appropriate sensor and analytical methodology, with sensors producing imagery at a

wide range of spatial, spectral and temporal resolutions (Mumby & Edwards 2002).

Habitat mapping commonly uses moderate resolution optical sensors that provide

sufficient detail to identify habitat boundaries at a scale relevant to management, while

being relatively cost effective and computationally efficient (Mumby et al. 2004).

Landsat TM is one of the most widely used satellite borne remote sensing technologies

for habitat mapping (Mumby et al. 1997). Landsat TM captures images at a spatial

resolution of 30m, has a return time of 18 days and includes 7 bands ranging from 0.42

µm (blue) to 2.35 µm (middle infrared). However, habitat mapping using this technology

requires turbidity levels to be low and regions to be cloud free. This can be problematic

in some tropical environments, where rainfall can be high and turbidity levels variable.

The north coast of Timor-Leste is recognised a relatively clear water environment, while

the south coast commonly experiences higher turbidity. The relatively high repeat

coverage of Landsat further provides opportunities to select imagery in which conditions

for habitat mapping are optimised.

Automated or semi-automated broad scale habitat mapping using remotely sensed

technology has often had difficulty producing cartographically acceptable products. This

is due to the ‘salt & pepper’ effect of pixel based classification. Object based image

analysis has been shown to produce classification outputs that are more

cartographically correct as it works within a framework that is based on the definition of

boundaries and identify patterns occurring at hierarchy of scale (Benz et al. 2004).

OBIA has been successfully applied for a number of land cover mapping applications

(Arroyo et al. 2006; Bock et al. 2005; Hajek 2006) although application to marine

environments has been limited.

Page 16 of 68

19

1.3 Project Aims & Objectives

This project’s primary aim was to develop a broad-scale habitat map of marine and

coastal habitats of the Timor-Leste northern coastline (approximately 250km) to assist

with future coastal-marine planning, biodiversity conservation and sustainable resource

use in Timor Leste. Specifically, this primary aim included (a) using Landsat TM

satellite imagery and object-based image analysis techniques to map broad-scale

habitats of Timor Leste; (b) undertaking fine-scale mapping of mangrove communities

by manual interpretation of aerial photography; (c) undertaking field ground-truthing or

validation via intensive field surveys of the marine environment and coastal habitats;

and (d) where appropriate, describing aspects of the composition, structure and

distribution of coastal habitats.

In undertaking this exercise, the project recognised the need to build on existing efforts

in localised, fine-scale mapping of coastal habitats (undertaken by MAF), and also, to

specifically engage and train MAF staff in coastal-marine habitat mapping and field

surveys.

2 PHYSICAL SETTING

Timor-Leste is located between Indonesia to the west and Papua New Guinea to the

east, and bounded by the Banda and Timor Seas to the north and south respectively

(Sandlund et al., 2001). Timor-Leste also includes the enclave of Oecussi and the

islands of Ataúro and Jaco. It has an area of 14,874 km2 with a coastline of more than

700 km. The island of Timor is part of the Banda Arc. The Australian continental crust

extends as far north as the north coast of Timor, and is thought to be uplifting Timor.

The Ramelau mountain range divides the country into north and south with its highest

peak reaching 2,963 m ASL (above sea level) (UNDP & RDTL, 2007). Over 40% of the

country has extremely steep slopes of 40% that are vulnerable to erosion and

constantly being worn down by the monsoonal rains, with numerous rivers draining to

the seas to the north and south (UNDP & RDTL, 2006).

Page 17 of 68

20

Figure 1: Study area identifying towns and villages referred to within this report.

The project study area stretched along the north coast of Timor Leste from the western

edge of Dili to south of Jaco Island (see Figure 1). The north coast of Timor Leste

experiences a dry tropical climate with a mean temperature above 24 °C and is

influenced by the Northern Monomodal Rainfall Pattern which sees a single wet season

from December to May. Annual rainfall in the north coast lowlands can be as low as

<1000 mm, whereas higher altitudes might receive rainfall up to 2000 mm/year.

Downpours are often extremely heavy (Barnett et al., 2003). In contrast, the southern

coast is exposed to two wet seasons (Nov-Apr, May-Jul) and around 1500mm of rain

annually.

The northern and southern coast differ not only climatically, but also, with respect to

coastal and nearshore environments. Topographically, the north coast is rocky and

steep along most of its shoreline. The continental shelf is narrow, with coastal plains

virtually non-existent or very narrow, except for around areas such as Manatuto and Dili,

with numerous white sandy beaches with interspersed rocky outcrops are scattered

along the coast. The north coast is characterised by karst geology and uplifted ancient

coral reefs (see Audley-Charles, 2004; Hamson, 2004; Keep, et al., 2009). The Page 18 of 68

21

dominant vegetation type along most of the coast is arid woodland (RDTL & CDU,

2006). In contrast, the continental shelf along the south coast is wide and gently

sloping, with relatively wide coastal plains characterised by river deltas, lagoons,

floodplains and swamps (RDTL and CDU 2006).

Along the northern coastline narrow fringing reefs and coral reefs can be found, as well

as seagrass meadows. The sheltered waters on the north coast, adjacent to the Banda

Sea (or Tasi Fetu) are calmer, deeper and clearer than on the south coast. In contrast,

the exposed waters of the south coast, adjacent to the Timor Sea (or Tasi Mane) are

rougher, shallower and more turbid – resulting in long stretches of sandy beach with

heavy waves and surf (Sandlund et al. 2001). As opposed to the low profile, extensive

coastal margin and plains of the southern coast, the northern nearshore littoral zone is

steep and very narrow - with the sea floor on the north coast dropping off sharply into a

3 km deep marine trench at approximately 20 km from shore (RDTL & CDU, 2006;

Keep et al., 2009).

Sandlund et al. (2001) identified the following biotopes in the coastal zone of Timor-

Leste:

Oceanic and sub-tidal marine environment: includes pelagic water columns,

deep-seabottoms, shallower rocky bottoms, sandy-muddy bottoms, seagrass

beds and coral reefs.

Intertidal zone: includes rocky intertidal shelves, sandy-muddy tidal flats and

mangrove forests.

Shorelines: include sandy beaches, dunes, rocky outcrops, limestone cliffs, river

estuaries, and brackish lagoons.

Near-shore zone: includes coastal drylands, natural forests and wetlands.

The coastal zone (and habitats) of Timor Leste are subject to a high degree of human

dependency and resource use. To-date, almost 560,000 people (approximately two-

thirds of the total population) in Timor Leste live in coastal and lowland areas with an

elevation up to 500 m ASL (NSD, 2006). Approximately 64% of the rural population is

food insecure, relying heavily on natural resources, with agriculture and (semi-)

subsistence fisheries being the major sources of income (~94%) for the population.

Most fishing activities are limited to low-technology inshore fishing. Mainly women and

Page 19 of 68

22

children collect fishes, crabs and molluscs in the intertidal zone at low tide (Sandlund et

al., 2001). During the Indonesian occupation, destructive fishing practices such as the

use of dynamite and cyanide were common. These days, such practices are still

sometimes used in some places (de Carvalho et al., 2007). Firewood is the primary

source of energy for 98% of the population (WB & ADB, 2007) and is widely collected

and logged.

3 METHODS

3.1 Input Data Two Landsat TM Scenes covering Timor Leste were acquired for the mapping of

marine and coastal habitats (see Figure 2). The images are Landsat 5 TM Scene

109/66 acquired on the 21st August 2004 and Landsat 5 TM Scene 110/66 acquired on

the 3 September, 2006. These images were selected based on; (1) tidal information for

Dili, Com, Lore and Suai that indicated that these date/times were low tide images (see

Figure 3); (2) late dry season images as this was identified as the preferable time to

optimise water clarity (3) low cloud cover, with an emphasis on marine/coastal areas;

(4) no scan line corrector (SLC) image errors and (5) captured within the 3 years of the

field program.

Figure 2: Landsat TM scenes covering Timor Leste.

Page 20 of 68

23Page 21 of 68

0

0.5

1

1.5

2

2.5

2/7/2003

17/7/2003

2/8/2003

18/8/2003

13/10/2003

16/5/2004

24/5/2004

17/6/2004

25/6/2004

5/8/2004

12/8/2004

20/8/2004

21/8/2004

5/9/2004

3/5/2005

19/5/2005

20/6/2005

7/8/2005

31/8/2005

8/9/2005

9/9/2005

24/9/2005

7/2/2006

23/2/2006

11/3/2006

12/4/2006

28/4/2006

14/5/2006

30/5/2006

15/6/2006

1/7/2006

17/7/2006

2/8/2006

18/8/2006

3/9/2006

19/9/2006

5/10/2006

21/10/2006

6/11/2006

22/11/2006

8/12/2006

24/12/2006

Date

Tide

hei

ght (

m)

Figure 3: Tidal heights for each date of Landsat image capture between 2003 and 2006. Ancillary datasets collected for mapping the marine and coastal habitats of Timor Leste

included the SRTM 3 second DEM, 1:50k aerial photography, coastline and road shape

files.

3.2 Image Pre-processing The Landsat images were pre-processed, to ensure that they were radiometrically and

geometrically correct. That is, the brightness value was corrected for errors caused by

changes in scene illumination, atmospheric conditions, instrument response

characteristics and viewing geometry. A simple black body subtraction correction was

applied in Erdas Imagine using deep water points located on the imagery as the images

were not being used for multitemporal analysis in this project. and the role of

radiometric differences between images was of less importance. GeoImage® supplied

the imagery as geometrically corrected (orthorectified using a cubic convolution

interpolation method). The spatial accuracy of this product was considered adequate

after comparison with existing mapping for Timor Leste.

Coastal areas of the imagery were delineated and selected by first defining the inland

extent based on the 30 m contour, as derived from the 3 second DEM. The seaward

extent was arbitrarily defined as 2.5 km from the coast (although the mappable area

due to water depth rarely was within a maximum of approximately 1km). The study area

Acquired Image

24

was created as a shape file in ArcGIS and used to mask the Landsat imagery in Erdas

Imagine. Following the creation of the input image layers, subsets were created based

on 6 training areas, each containing approximately 10km of coastline. These included

Metinaro, Manatuto, Baucau, Com, Jaco and Lore. These were used to investigate the

potential for object based image analysis to map the marine and coastal habitats of

Timor Leste. Object based image analysis can be processing intensive and this

approach allowed different segmentation and classification approaches to be trialed and

assessed before the more lengthy process of application to the fully study area (see

Figure 4).

Figure 4: Subset focus regions used for the development of the mapping approach.

Finally, a number of image products were created to assist in the classification process.

The 3 second DEM was used to calculate a slope layer using a 3 x 3 neighbourhood as

defined in ArcGIS 9.2 while principle components, band combinations (e.g. Normalised

Difference Vegetation Index - N DVI) and depth invariant indices were calculated from

the Landsat imagery using ENVI or Erdas Imagine software. Principle components

analysis was used to produce a new set of image layers that may better capture the

information contained in the original imagery. That is, some image layers are often

highly correlated, meaning that the two layers duplicate a very similar set of information.

Principle components analysis transformed the information in correlated layers into a

smaller number of layers, which are referred to as the principle components (PC). The

Page 22 of 68

25

first and second PC contains the greatest variation in the data with subsequent principle

components generally being of less value. Band ratios have also been used to highlight

different aspects of the environment captured in remotely sensed imagery. The most

frequently used NDVI which is defined as ρNIR-ρRed)/( ρNIR+ρRed) where ρNIR and

ρRed are reflectance values of near infrared and red light. This index ranges between -

1 to 1, with highly photosynthetic vegetation approaching 1. The NDVI was calculated to

assist with classifying coastal vegetation. Finally, depth invariant layers were created to

assist with the marine habitat classification. This was based on the methods described

in Lyzenga (Lyzenga 1978; Lyzenga 1981) which produces a single depth invariant

band from each pair of spectral bands (Mumby et al. 1998) using pixel values identified

for a similar substrate across a variety of depths. This was obtained for predominantly

sandy substrates.

3.3 Image Classification

3.3.1 Pixel Based A preliminary image classification was undertaken using an unsupervised classification

technique. The isoclass algorithm was applied in Erdas Imagine with 50 classes created

(see Figure 5).

Figure 5: Unsupervised classification of the Landsat imagery in the Manatuto study region.

These were then amalgamated to produce broad classes and were used to identify the

spectral separability of different classes. This was applied to all areas and to marine

and coastal areas separately. The classifications indicated better class separability Page 23 of 68

26

within the terrestrial habitat classes relative to marine habitats. However, the highly

pixilated output suggested that object based methods would be more appropriate for

the production of a final vector based mapping product.

3.3.2 Object Based Image Analysis Image Segmentation

Object based classification requires that imagery is segmented prior to classification.

The segmentation process is a crucial component for successful classification of the

imagery as it is intended to create homogeneous object primitives that are

representative of features within the image. These image object primitives are then

used as information carriers and building blocks for classifying the relevant features

within the image (Definiens 2006)

The segmentation process requires various parameters to be set, such as the scale

parameter, the single layer weights, and the mixing of heterogeneity criterion

concerning tone and shape. These parameters are user-defined and are set

accordingly with consideration to the requirements of the analysis, commonly through a

process of ‘trial and error’. The scale parameter defines the maximum allowed

heterogeneity for the resulting image objects and has the greatest influence on the size

of objects created during segmentation. The homogeneity criterion determines the

weighting between colour and shape, with shape further divided by smoothness and

compactness. The user can therefore determine the influence that object shape has on

the segmentation process, independent of image spectral properties. This is particularly

useful when detecting objects with a common shape characteristic.

Image segmentation can be performed using all or a weighted selection of input layers

and can also be undertaken at a range of scales to create a object hierarchy. That is,

image layers considered more important in the initial segmentation process can be

more heavily weighted to better identify appropriate object boundaries (for example an

NDVI layer may be given twice the weighting of individual bands for vegetation

mapping). This can be varied depending on the scale of mapping. For example large

objects may be created to initially stratify the landscape and is likely to undertaken

using equal weighting of input layers. While a fine scale segmentation might be

Page 24 of 68

27

performed only using one layer (e.g. NDVI) to identify fine scale variations in vegetation

cover. The larger and smaller objects will be nested and share common boundaries

allowing a hierarchical approach to the classification to be undertaken.

The subset regions were used to examine segmentation approaches for marine and

coastal habitat mapping in Timor-Leste. Various segmentation parameters were

applied, with the final approach deemed most suitable for classification based on the

creation of two segmentation layers. A scale value of 10 produced objects at a scale

deemed most suitable for final habitat mapping, while finer scale objects produced

using a scale parameter value of 5 was found to identify smaller objects useful in the

final classification. Different homogeneity criterion parameter values were examined;

however, the default values tended to produce the most suitable results and were

applied to the final segmentation.

A single segmentation and classification approach was used to map both the marine

and coastal habitats, as this provided a seamless transition in the mapping of the two

environments (see Figure 6). This approach, however, caused the indices that

highlighted different aspects of the terrestrial (e.g. NDVI) and marine (e.g. depth

invariant bands) habitats were not applicable across both environments. As such, the

optimal segmentation results were obtained by equally weighting the original imagery.

Figure 6: Segmented objects for northern Jaco island for scale parameter value 10 (a) and scale parameter value 5 (b) (Landsat imagery displayed as true colour).

Page 25 of 68

28

Image Classification

The classification process in Definiens object based image analysis requires the initial

establishment of a class hierarchy, followed by the identification of the unique spectral,

textural and spatial attributes of each class and finally the selection of algorithm for

allocating objects to a particular class. The class hierarchy established for mapping

Timor-Leste coastal and marine habitats included 22 classes based on visual

interpretation of the imagery, and were refined to 14 classes based on comparison with

field data.

Figure 7: Definiens classification environment showing the class hierarchy developed within this project and example class description of ‘Mangroves’.

The classification made use of a number of object features object that distinguish the

habitat class (see Figure 7). For example, while the bare sand class was based solely

on the spectral characteristics of this class in bands 1 to 5, the mangrove class used a

combination of spectral information (to identify it as relatively highly photosynthetic

vegetation) and spatial properties (to identify that it would be adjacent either to another

polygon classified as mangroves or a marine habitat class). The rule set developed

using these features allocated objects to specific classes using a combination of the

Page 26 of 68

29

standard nearest neighbour algorithm, based on a set of defined training polygons, or

membership function. Membership functions allow a fuzzy rule base to be developed,

requiring the definition of a membership function shape and estimated minimum and

maximum cut-off values. These can be highly valuable for capturing expert knowledge

of specific mapping classes and their characteristics.

3.3.3 Manual Interpretation Fine-scale (1:50,000) mapping of mangroves on the northern coast of Timor-Leste was

undertaken through expert interpretation of the 1:50,000 2002 aerial photography and

supported by inspection of imagery available through Google Earth® (generally very

high resolution ‘QuickBird’ satellite imagery) (see Figure 8). The mapping was

undertaken in ArcGIS 9.2, with attributes including identification of the community,,

dominant vegetation type, general description, environmental setting and area recorded

for each polygon. A total of 18 communities could be identified from the imagery based

on the spatial, spectral and textural properties of the imagery and comparison with field

based observations.

Sonneratia alba/ Rhizophora Sonneratia alba

Salt Flat

Ceriops with Avicennia/Lumnitzera

Figure 8: Aerial photography of a mangrove/salt flat sequence showing some of the dominant vegetation communities identified through manual interpretation.

Page 27 of 68

30

3.3.4 Fine-scale Mangrove Mapping (1:10,000) Finer scale mangrove mapping was also undertaken (by NT Government) as an adjunct

to the Landsat mapping, with mangrove communities mapped through interpretation of

aerial photography at the 1:10,000 scale.

Areas mapped to-date include only the north coast of Timor Leste from the West Timor

border through to the east tip of Timor Leste at Tatuala. Spatial resolution of the final

mapping is at the 1:25,000 scale. However, further ground-truthing and validation and

‘on-site’ checking is required.

3.4 Coastal-Marine Field Surveys

3.4.1 Coastal Habitats

A vegetation survey covering the coastal habitats of the Timor-Leste north coast was

undertaken in November 2007 (see Appendix 3). Each of the subset study regions (see

Figure 4) were surveyed to ensure that most habitat classes were represented. Within

each region, approximately 20 points were surveyed with over 100 data points collected

(see Figure 9). The survey was based on an established method used within the

Northern Territory Government for ground-truthing vegetation mapping. Information

collected included approximations of the percent canopy cover based on a 10m

diameter field of view and height of the ground, mid and upper vegetation layers. The

dominant species (with a maximum of 3) were recorded in the mid and upper layers.

Information was also recorded about the substrate characteristics and general

environmental setting. Site locations were collected using a ProMark3 GPS with an

accuracy of 5-7 m.

Page 28 of 68

31

Figure 9: Mangrove forest surveyed in Manatuto region showing points collected, vegetation structural classification and Landsat imagery.

The data recorded was classified based on the National Vegetation Information System

(NVIS) of Australia. The NVIS structural classification is built from the following

attributes derived from vegetation site data:

• Growth form (for each stratum);

• Cover (%) value (for each stratum);

• Height (m) range and average (for each stratum);

• Floristics (up to 5 dominant species for each stratum), and

• Basal counts (to assist in determining dominant woody species).

The site data was classified into structural classes (based on all structural classes and

by the upper canopy structure only) and floristic groupings.

3.4.2 Marine Habitats

Study Sites

Six regions were selected across the north, east and south-east coast of Timor-Leste at

Metinaro, Manatuto, Baucau, Com, Tutuala and Jaco Island, and Lore (see Figure 4).

These regions were chosen for ground-truthing remote sensing imagery (Landsat 7,

Page 29 of 68

32

30m x 30m pixel size) as they contained representative groups of habitat types and

geomorphology, and also, provided access to community infrastructure and

opportunities for capacity-building. Com, Tutuala, Jaco Island and Lore are also located

within the proposed marine park adjacent to the newly declared Parque Nacional Nino

Koni Santana (Nino Konis Santana National Park). Community-level consultations with

Suco village chiefs, regional MAF officers and local fishermen, were also undertaken

prior to initiating fieldwork (see Plate 1).

Sampling

Sampling was conducted from 20th to 30th November 2007 during the late dry season

by paired staff from Ministry for Agriculture and Fisheries (MAF) and the Northern

Territory Government (NTG). Between two to four local fishermen and their boats were

used at each region. Transects were selected to optimize coverage of predicted habitat

types and boundaries within a region, and were generally orientated perpendicular to

the coastline (see Figure 11). Over 700 random points were sampled along transects,

with increasing frequency near transitions of habitat type or geomorphology. Sampling

was undertaken using improvised bathyscopes and drop down video cameras (see

Figure. 10). Site identification, video start and finish times, depth, latitude, longitude,

geomorphology and dominant habitat classification were recorded on tabulated sheets

(Appendix 1), data was then validated and entered into a Microsoft™ Access database

(Appendix 2).

Figure 10: Sampling equipment used in the marine survey, including improvised bathyscopes and drop-down video cameras.

Page 30 of 68

33Page 31 of 68

Plate 1: (a) Training of MAF staff in marine habitat identification and use of underwater habitat mapping technology; (b) community meeting and consultation (Lore); (c)-(h) use of local boats and fishermen to undertake habitat mapping.

(a)

(c)

(e)

(h)

(f) (g)

(b)

(d)

34Page 32 of 68

Figu

re 1

1: E

xam

ple

of m

arin

e ‘g

roun

d-tru

thin

g’ sa

mpl

ing

sites

and

cro

ss-s

helf

trans

ects

in th

e M

anat

uto

regi

on.

35

Classification

Geomorphology classification was adapted from the generalized diagram of a coral reef

cross section by Mather and Bennett (1994) and incorporates a secondary classification

for rugosity (relief or height of corals), coral patch density, and lagoon depth (see

Appendix 1).

Benthic habitat classification was adapted from English et al. (1997). Habitats were

broadly classified according to the dominance or absence of biota, such as coral, algae,

seagrass, and bare substratum (Appendix 1). Secondary classification of habitat

incorporated additional descriptions of coral morphology (life form), algae colour or

type, sediment type, and seagrass density. Notes on reef condition were collected for

anthropogenic impacts (dynamite, anchors) and general health (bleaching, nutrient

indicator algae, crown of thorns, and bleaching).

3.4.3 Classification Refinement, Validation & Output The Landsat derived mapping of coastal and marine habitats on the Timor-Leste north

coast was compared to the field survey data to refine and validate the RS based

classification. That is, the mapping and field survey classification schemes differed

based on the scale of data collection and methodology used by each approach. A

matrix comparing the mapped with ground derived classes was created. The RS

classification system was developed based on refining points of overlap and duplication

within RS classes and identification of field based classes that could not be separated

at the scale of mapping. The associated field derived vegetation and marine habitats

contained within each mapped habitat class were identified and included in meta data

information associated with the final map.

The final map was also processed within ArcGIS to improve the classification, ensure

topological integrity of the final map and to produce a more cartographically acceptable

product by smoothing the output polygon boundaries. Manual classification assisted in

both correctly classifying misclassified objects and ensuring consistency across the

dataset to better represent continuous habitat types. The final classification was then

simplified using the simplify polygon tool to remove extraneous bends and small

intrusions and extrusions from a line without destroying its essential shape (ESRI

Page 33 of 68

36

1996). Finally, adjoining boundaries between polygons classified as single habitat and

overlapping polygons classified twice were cleaned using the dissolve tool in ArcGIS.

4 RESULTS Based on the results of remote sensing analyses and the coastal-marine “field ground-

truthing”, a broad scale (1:100k) map of the coastal and marine habitats of the north

and east coast of Timor-Leste has been produced for approximately 250km of the

Timor-Leste coast line – covering the entire northern coast to, south of, and including,

Jaco Island. Examples of the mapping have been provided for six regions of the coast

in Figures 12 and 13. This dataset has been produced using object-based image

analysis techniques and has been prepared for cartographic presentation. The map is a

topologically correct, GIS ready data layer and includes three attribute fields including

the ‘Class’ (a descriptive name of the habitat), ‘ClassCode’ (a unique code for the

habitat type) and ‘Area_ha’ (the area of each polygon. The smallest mapped element is

0.1ha, largest is 1040 ha with a mean polygon size of 10ha (excluding the ‘Deep Water’

class). As expected, the dataset is highly linear, having a maximum ‘width’ of

approximately 3km, with much of the mapping being less than 1km in width.

The classification identified a total of 6 marine habitat types and 7 coastal habitat types

(see Plate 2). The marine nearshore zone is characterised by a narrow reef flat (often <

60m wide, but up to almost 1km), dominated by seagrass in shallower water

(approximately 2,200ha) and corals in deeper water and on the escarpment

(approximately 2,000ha). A mixed-cover class, which included low covers of coral or

seagrass and bare areas, occupied 1,250 ha. Within the coastal habitats, mangroves

occupied 750ha, the largest stand being in the Metinaro region. Detailed mangrove

mapping (1:10,000) identified a total of 20 mangrove communities with clear zonation

patterns. Bare areas (including salt flats and beaches) occupied 1,100ha, while the

vegetation was dominated by savannas, with 3,750ha of open woodland or shrubland,

1,935ha of open forest or woodland and 1,050ha of coastal forest. Intensive agriculture

(2,200ha) and built up areas (1730ha) were also a prominent feature of the northern

coastline of Timor Leste.

Page 34 of 68

37Page 35 of 68

Plate 2: Coastal and marine habitats along the north coast of Timor Leste; (a)-(b) small stands of mangroves; (c)-(d) fringing coral reef flats; (e) limestone reef pavement (Com); (f) pocket beaches (One Dollar Beach); (g) fringing mangroves (Lore); (h) beach (Tutuala).

(d)

(f)

(g)

(h)

(b)

(a)

(c)

(e)

38

a) b)

d)c)

f)e)

Figure 12: Coastal and marine habitat mapping in the (a) Jaco, (b) Dili, (c) Bacau, (d) Manatuto, (e) Com and, (f) Metinaro, regions

Page 36 of 68

39

Figure 13: Map of the coastal and marine habitats in the Jaco Island region. [Insert shows the

extent of broadscale coastal and marine habitat mapping conducted along the north coast of

Timor Leste.]Page 37 of 68

40

The final habitat classes mapped and their distribution are discussed in more detail in

the following sections.

4.1 Coastal Habitats

4.1.1 Overview of Coastal Habitats

The seven coastal habitat classes included in the final mapping of the Timor-Leste

north coast are shown in Table 1. The field survey identified 14 structural classes

(based on the structure of the upper strata), 45 classes based on all strata and 24

floristic groups. The vegetation ranged from closed forest dominated by mangrove

species (e.g. Avicennia marina, Rhizophora stylosa and Sonneratia alba), through

palmlands and woodlands dominated by Borassus flabellifer and Tamarindis indica

respectively to bare salt flats with a total of 42 dominant plant species identified

(Appendix 4).

Habitat Code Area (ha) Bare Areas including Beaches, Salt Flats and River Channels

SF 1108

Coastal Forest (Non Mangrove) CF 1061 Mangrove Coastal Forest CFM 754 Open Forest or Woodland W 1935 Open Woodland or Shrubland OWS 3750 Agriculture (Rice Paddies) RP 2203 Built Up (Urban and Semi Urban) Areas U 1732

Table 1: Coastal habitat classes mapped and their total area.

4.1.2 Beaches & Saltflats

Bare areas mapped included salt flats, beaches and river channels (see Figure 14).

These occupied approximately 1100ha. The largest continuously bare areas, up to 115

ha, occurred on river channels and deltas (see Table 2). Salt flats occupied areas up to

30ha and often adjoined areas of mangroves. Beaches were somewhat inconsistently

mapped due to the resolution of the imagery. However, the longest beach mapped

stretched was approximately 10km with an average width of nearly 30m (1 pixel). Field

survey indicated that the bare areas included a range of substrate types, ranging from

mud to coarse sand and pebble beaches.

Page 38 of 68

41

Class Number of Polygons

MaximumArea (ha)

Average Area (ha)

Standard Deviation

Agriculture (Rice Paddies) 27 401 81.6 108.7Bare Areas including Beaches, Salt Flats and River Channels

140 115 7.9 15.9

Built Up (Urban and Semi Urban) Areas 17 1040 101.9 252.7Coastal Forest (Non Mangrove) 60 98 16.2 23.3Mangrove Coastal Forest 42 304 20.0 50.1Open Forest or Woodland 267 182 7.2 17.5Open Woodland or Shrubland 287 663 13.1 46.2

Table 2: Size and number of polygons mapped within each coastal habitat class.

Figure 14: Salt flat (a) and beach (b) on the north coast of Timor Leste.

4.1.3 Coastal Forests (non-mangrove)

Coastal forests (non-mangrove) occupied approximately 1000ha and occurred in

patches of up to 100ha but were on average 16 ha in area (see Figure 15). The field

survey indicated that these mapped areas were variable and included both open forest

and closed palmland structural groupings dominated by Corypha utan, Tamarindis

indica, Borassus flabellifer and Schleichera oleosa.

Page 39 of 68

42

Figure 15: An example of coastal forest on the Timor-Leste north coast.

4.1.4 Mangroves

The north coast of Timor-Leste supports approximately 750ha of mangroves, with more

than half of these occurring in an almost unbroken stretch approximately 12 km long in

the Metinaro region (see Figure 16). Seven species of mangrove were identified and

the field survey indicated a clearly defined pattern of zonation, typical in mangroves.

Detailed mapping from the aerial photography identified approximately 20 communities

based on structural and floristics characteristics (see Figure 17, Table 3). The

mangrove communities occupying the greatest area includes Rhizophora dominated

closed forest (+/- Sonneratia), Sonneratia alba closed/open forest (+/- Rhizophora),

Sonneratia alba closed forest and Ceriops dominated low closed forest (+/-

Avicennia/Lumnitzera).

a) b)

Figure 16: Examples of mangrove forest a) Rhizophora dominated closed forest and Ceriops

dominated low closed forest.

Page 40 of 68

43

The results from the fine-scale mangrove mapping (1:10,000) is summarised in Table 3.

More detail will be published in a separate scientific report and publication at a later

date.

Dominant Mangrove Community Area(ha)

Sonneratia alba-Rhizophora dominated forest 285Saltflat/saltpan 176Sonneratia alba dominated forest 103Hinterland 101Ceriops dominated +/- Avicennia+/-Lumnitzera forest 53Mangrove/palm forest 39Mixed species Ceriops/Rhizophora/Avicennia forest 26Avicennia marina dominated forests 26Lumnitzera racemosa/Avicennia marina +/- Excoecaria low open-forest 17Lumnitzera racemosa/Avicennia marina +/- Sonneratia alba low open-forest 14Swamp shrubland 12Unknown 11Ceriops dominated forest 8Rhizophora stylosa/Avicennia marina low open forest 7Mixed Avicennia marina/Ceriops 7Avicennia marina/ Rhizophora stylosa open forest 5Rhizphora apiculata/Sonneratia caseolaris/Aegiceras corniculatum low open-forest/woodland 3Strand or beach 3Ceriops australis sparse low woodland 2Lumnitzera/Azima mid spase shrubland-saltpan 1Terrestrial 1Total: North Coast Timor Leste 899

Table 3: Mangrove communities (and extent) recorded along the north coast of Timor Leste (West Timor border to Tutuala).

Page 41 of 68

44

Figure 17: Example of the fine-scale (1:10,000), mangrove community mapping undertaken.

4.1.5 Open Forest, Woodland & Shrubland

Open forest or woodland and open woodland or shrubland were the most extensive

habitat classes mapped (approximately 6000ha) and included the greatest number of

vegetation communities identified in the field surveys (see Figure 18). Although it was

difficult to define boundaries with non-coastal habitats, some key species appeared to

dominate the coastal zone. In particular, Tamarindis indica, Zyzyphus mauritiana and

Borassus flabellifer were key species in the upper strata.

Page 42 of 68

45

a) b)

Figure 18: Examples of a) open woodland or shrubland and b) open forest or woodland habitat

classes.

Agriculture was mapped in 6 regions and covered more than 2000ha in the coastal

zone, with large areas around Manatuto, Laleia/Vernasse and Laga (see Figure 19).

These areas were predominantly growing rice and occupied floodplains adjacent to

larger rivers in the region. Built up areas were also mapped, with approx. 1700ha

shown. However, this mapping primarily included larger built up areas and some small

settlements may have been missed due to the scale of the imagery.

Figure 19: Rice paddies at different stages of production.

4.1.6 Coastal Mapping Accuracy

Comparison of the field survey data with the coastal habitat mapping reveals broad

agreement when the field survey data are grouped along similar lines with an overall

accuracy of 83% (Table 4). Mangroves are mapped well, with commission and omission

Page 43 of 68

46

accuracies of 93% and 90% respectively. This habitat class is very well sampled

(almost half the points recorded within this habitat), as these data were also used to

guide the fine scale mangrove community mapping. Open woodland or shrubland and

woodland classes are also reasonable well sampled, with the woodland class achieving

higher mapping accuracies. The open woodland or shrubland class had some

confusion at the boundaries of this class, tending to overlap with other classes. Bare

areas surprisingly recorded the lowest accuracy. This may be a function of the Landsat

pixel resolution, with the full pixel required to be completely devoid of vegetation to be

classed as a bare area, while as the salt flats were often small and points surveyed to

close to vegetation (see Figure 20). Spatial errors also caused three coastal habitat

points to be mapped as marine habitats. Coastal forests and rice paddies were not

surveyed sufficiently. However, the two datasets were in agreement and further

inspection of aerial photography indicated that these two habitat classes were

reasonable well mapped. Further information on the floristic and structural vegetation

classes (based on the field survey) contained within each mapped class can be found in

Appendix 5.

Field Survey CF CFM OWS RP SF W TOTAL Commission

CF 2 0 0 0 0 0 2 NACFM 0 37 0 0 2 1 40 93% DW 0 1 0 0 0 0 1 0% OWS 0 1 14 0 2 1 18 78% RFSL 0 1 0 0 1 0 2 0% RP 0 0 0 1 0 0 1 NASF 0 1 2 0 5 0 8 63% W 0 0 1 0 0 10 11 91% TOTAL 2 41 17 1 10 12 69

Land

sat M

appi

ng

Omission NA 90% 82% NA 50% 83% Overall 83%

Table 4: Accuracy assessment for the coastal habitat mapping.

Page 44 of 68

47

Field ‘nil upper’ point. Mapped as CFM

Figure 20: Example of a ‘nil upper’ point survey within close proximity to vegetated land cover.

4.2 Marine Habitats The marine habitats mapped from Landsat included 6 classes that represented different

groupings of different geomorphic and benthic characteristics identified on the Timor-

Leste north coast (Table 5). Previous studies have found that Landsat can provide

reasonable estimates of coarse descriptive resolution habitat classes (e.g. sand, coral,

seagrass, algae) based primarily on the spectral properties of different benthos (Mumby

et al. 1997). The final classification provided here has a similar definition of benthos,

with high and low seagrass separated and a spatial element included to separate a fore

reef and escarpment class. Mixed coral, seagrass and open reef flat tended to

represent areas where cover was very low and the substrate was predominantly bare.

Habitat Code Area (ha)Coral Dominated Fore-Reef and Escarpment FRC 1558 Coral Dominated Reef Flat RFC 458 Deep Water, with coral likely in shallower areas DW 62708 Dense Seagrass Covered Reef Flat RFSH 933 Mixed Coral, Seagrass and Open Reef Flat RFM 1266 Sparse Seagrass Covered Reef Flat RFSL 1173

Table 5: Marine habitats and their area mapped on the Timor Leste north coast.

The area of each class mapped on the Timor-Leste north coast is also shown in Table

5, while the size distribution of habitat patches are shown in Table 6. Excluding deep

Page 45 of 68

48

water, coral dominated fore reef and escarpment occupied the largest area of habitat

mapped. This habitat class occurs along most of the Timor-Leste north coast and is the

dominant habitat class where there is no coastal plain and the mountains plunge rapidly

into the Banda Sea (see Figure 21). However, the area of this class may be

overestimated as the definition of the seaward extent of the habitat was influenced by

water clarity. That is, the class occurs in deeper water and a small increase in turbidity

may cause the object to be assigned to this class. This was compensated for by

including a spatial expression in the classification (e.g. adjoining a reef flat class) and

was further edited in the manual editing phase.

Figure 21: Example of narrow reef with sharp drop off and associated Landsat TM mapping (note imagery is aerial photography).

The coral dominated reef flat occupied approximately 460ha and was generally located

on the outer reef flat (see Figure 22). The class tended to overlap with the coral

dominated fore reef and escarpment, particularly where the reef flat was narrow, due to

the resolution of the imagery. Results of the field survey supported this transition, with

coral observed at a minimum depth of 1.1 m and average depth of almost 1.6m within

Page 46 of 68

49

the reef flat geomorphic class. Although this class could be found across the northern

Timor-Leste coastline, larger areas occurred within the 40km of coastline to the east of

Dili and around Com to Jaco Island in the east.

Figure 22: a) Lagoon and Reef Flat complex east of Dili and b) coral dominated reef.

Class Number

of Polygons

Maximum Area (ha)

Average Area (ha)

Standard Deviation

Coral Dominated Fore-Reef & Escarpment 154 97 10.1 14.7Coral Dominated Reef Flat 119 53 3.8 7.6Deep Water, with coral likely in shallower areas 13 62,641 4,823.7 17,371.9Dense Seagrass Covered Reef Flat 234 124 4.0 12.1Mixed Coral, Seagrass & Open Reef Flat 302 44 4.2 6.0Sparse Seagrass Covered Reef Flat 219 165 5.4 16.9Table 6: Size and number of polygons mapped within each marine habitat class.

Seagrass and coral were found to occupy a similar total area around the Timor-Leste

north coast near-shore environment (approximately 2,000ha) (see Figure 23). Seagrass

tended to dominate the shallow waters of the reef flats adjacent to the shore at an

average depth of 1.0m and minimum of 0m, with more dense seagrass predominantly

found in deeper water (> 1m). Within points classified as seagrass from the field survey,

the average seagrass cover was 43% and a maximum cover of 100% recorded. Low

seagrass covers were also recorded during the field survey some points classified as

coral and bare. The mapped mixed coral, seagrass and open class tended to include

Page 47 of 68

50

areas with low cover, with the presence of either seagrass or coral, or often neither

possible and often occurred between a coral and seagrass dominated habitats.

Figure 23: a): Exposed seagrass covered reef flat and b) high density seagrass cover observed

on the north coast.

A traditional validation of the mapping could not be undertaken due to differences in

class definition between the field survey and mapping. However, comparisons with the

dominant benthic cover and geomorphic classification could be made (Table 7 and

Figure 24). Overall, there was approx. 8% true error (92% accuracy) in classification

between the mapped data and field survey of dominant benthic cover (e.g. an area

mapped as coral but recorded in the field as seagrass). High seagrass cover reef flats

appeared to include the greatest error, with almost a third of the points within this

habitat being bare. This highlights the difficulty in defining a boundary between high and

low seagrass covers using Landsat.

Class Code Algae Bare Coral Seagrass Grand TotalDW 25 48 2 242FRC 20 112 1 172RFC 7 15 2 29RFM 29 34 20 88RFSH 2 18 3 40 66RFSL 20 2 21 48Grand Total 2 120 215 86 648

Table 7: Comparison of mapped classes with field survey dominant benthic cover (numbers are counts).

Page 48 of 68

51

Figure 24: Habitat map with marine survey data overlain in the Metinaro region (see Table 1 for marine habitat codes).

5 DISCUSSION

5.1 Coastal-Marine Mapping

Timor-Leste has a coastline of 706km, however to-date, research on the extent and

status of its’ coastal and marine habitats, biodiversity and marine resources, has been

limited. While marine research studies have been undertaken, they have predominantly

focussed on localised, coral reef surveys (Hodgson 1999, Deutsch 2003, Wong & Chou

2004, Dutra and Taboada 2005) rather than habitat mapping. The current project has

undertaken remote sensing analyses and coastal-marine field validation and ground-

truthing to produce the first, broad-scale (1:100k) map of the coastal and marine

habitats of the north and east coast of Timor-Leste.

The current project has also confirmed that coastal-marine habitats on the north coast

of Timor Leste are limited – with most of the coastal estate being highly linear and

narrow in extent (ie. approximately 3km in width, with much of mapping confined to less

Page 49 of 68

52

than 1km in width). As such, coastal physiography (particularly the steep coastal

gradient and the absence of significant coastal plains) has resulted in a coastline

characterised by steep coastal cliffs and rocky headlands, interspersed with pocket

beaches and narrow fringing reefs – with limited development of estuaries, mangrove

forests (750ha), seagrass meadows (2,200ha) and coral reefs (2,000ha). With

increasing pressure on coastal-marine habitats and their resources (for fuel, timber and

food) in Timor Leste (Sandland 2001) and their limited extent - these habitats are

indeed highly vulnerability to further loss and over-exploitation. This underscores the

need for precautionary, adaptive approaches to coastal development and also,

emphasises the need for urgent progress on implementing integrated coastal planning

and management in Timor Leste, including coastal protection, regulations and

rehabilitation measures.

5.2 Mangroves & Coastal Wetlands

5.2.1 Distribution Mangroves occupy only a small area of Timor Leste (1,802ha) when compared with

neighbouring regions such as Indonesia (3,062,300ha) and Australia (1,451,411ha)

(see Table 8). This is due mostly to the coastal configuration and physiography of Timor

Leste, which unlike other islands of the Indonesian archipelago and the north coast of

Australia, does not include the salient coastal contours, physiographic features and

coastal processes, for significant mangrove development (ie. extensive low-lying

coastal plains, sheltered waters, sedimentary processes). Further, rivers in Timor Leste

flow intermittently and hence, large estuarine systems are generally absent.

Mangroves generally predominate on the north coast in inlets with calmer, protected

waters, with the largest contiguous block of mangroves found in the Metinaro region. In

contrast, on the wave exposed south coast, mangroves are generally confined to the

mouths of the streams, and marshy or swampy regions.

Page 50 of 68

53

Country Area (ha) *1 % World mangroves

No of true mangrovespecies *2

Timor Leste 1,802 0.001 19 Northern Territory 373,700 (woody

only) 2.4 37

Papua New Guinea 410,000 2.7 37 Australia 1,451,411 10 37 Indonesia 3,062,300 19 43 *1 Figures from FAO 2007 and Timor Survey *2 Species showing high fidelity to mangrove communities: additional species found near or occassionally in mangrove communities do occur.

Table 8: Regional comparison of mangrove extent and diversity. Based on recent FAO analysis (FAO 2007) approximately 15.2 million hectares of

mangroves are estimated to exist worldwide circa 2005 (down from 18.8 million

hectares in 1980). The most extensive mangrove area is found in Asia, followed by

Africa and North and Central America. Five countries (Indonesia, Australia, Brazil,

Nigeria and Mexico) together account for 48% of the total global area, with 65% of the

total mangrove area found in just ten countries. The remaining 35 percent is spread

over 114 countries and areas, of which 60 have less than 10 000 ha of mangroves

each. The mangroves in Timor Leste (1,800 ha) account for approximately 0.001% of

the estimated global mangrove extent.

5.2.2 Species Diversity Although the extent of mangroves is small in Timor Leste, in comparison to

neighbouring regions, they are reasonably diverse (ie. 19 species). Most common

mangrove genera are represented, although taxa found are less diverse than

surrounding regions. Most species found in Timor are also found along the Northern

Territory coastline and in Indonesia. Community composition and zonation is also

similar to that found in the Northern Territory with the exception of Sonneratia spp.

dominated forest communities which are found on coralline benches on the seaward

margins in Timor Leste. These are not found in Northern Territory mangrove habitats.

These coralline benches are also at a higher tidal level than the adjacent tall

Rhizophora spp. dominated forest communities, the reverse of that occurring in the

Northern Territory.

A total of 19 mangrove species were found during this survey (Table 9) and this

information was used to update species information from previous surveys (ie. FAO Page 51 of 68

54

2007). More species however are likely to be included following further field survey

work. All species found in Timor Leste are found in neighbouring regions. Using current

survey results and results from previous surveys in the region (FAO 2007), Table 10

provides a comparative overview of the individual mangrove species found within the

region.

Mangrove species recorded during current 2007 survey Acanthus ilicifolius Heritiera littoralis Acrostichum auream Lumnitzera racemosa Aegiceras corniculatum Pemphis acidula Avicennia marina Rhizophora apiculata Bruguiera gymnorrhiza Rhizophora stylosa Bruguiera parvifolia Scyphiphora hydrophylacea Ceriops australis Sonneratia alba Ceriops decandra Sonneratia caseolaris Excoecaria agallocha Xylocarpus mekongensis Excoecaria ovalis

Table9: Species of mangrove recorded in the current coastal mapping and survey of Timor Leste.

Species Indo

nesi

a

Aus

tral

ia

Tim

or L

este

Pa

pua

New

Gui

nea

Com

mon

to

all R

egio

ns

Acanthus ebracteatus * * Acanthus ilicifolius * * * * * Acrostichum aureum * * Acrostichum speciosum * * Aegialitis annulata * * * Aegiceras corniculatum * * * * * Aegiceras floridum * Avicennia alba * * Avicennia eucalyptifolia * Avicennia integra * Avicennia marina * * * * * Avicennia officinalis * * Avicennia rumphiana * * * Bruguiera cylindrica * * * Bruguiera exaristata * * * Bruguiera gymnorrhiza * * * * Bruguiera hainesii * * Bruguiera parviflora * * * * * Bruguiera sexangula * * * Camptostemon philippinensis * Camptostemon schultzii * * *

Page 52 of 68

55

Ceriops australis * * * Ceriops decandra * * * * Ceriops tagal * * * Cynometra iripa * * Cynometra ramiflora * Excoecaria agallocha * * * * * Excoecaria indica * * Excoecaria ovalis * * Heritiera globosa * Heritiera littoralis * * * * * Kandelia candel * * Lumnitzera littorea * * * Lumnitzera x rosea * * * Lumnitzera racemosa * * * Nypa fruticans * * * Osbornia octodonta * * Pemphis acidula * * * * Rhizophora apiculata * * * * * Rhizophora mucronata * * * Rhizophora samoensis * Rhizophora stylosa * * * * Rhizophora x lamarckii * * Scyphiphora hyrophyllacea * * * Sonneratia alba * * * * * Sonneratia caseolaris * * * Sonneratia ovata * * Sonneratia x gulngai * * Sonneratia x urama * * Xylocarpus mekongensis * * * * * Total 43 37 19 37 9

Table 10: Comparison of mangrove distribution in Timor Leste and the region. Based on results from the current 2007 coastal survey and previous regional accounts (FAO 2007).

5.2.3 Community Structure Mangroves of Timor Leste show typical zonation in community structure as found

elsewhere in the region. Zonation is determined by tidal regime and substrate type.

Community structure varies from bare saltpans through low open woodlands, low

closed forests to taller open forests and closed forests. Species composition varies from

mono-specific stands to those with 4 or 5 species.

On the seaward margins where tidally inundation is regular, Sonneratia alba and

Bruguiera parvifolia appear commonly in the sandy sediments adjacent to beaches,

Rhizophora spp. and Bruguiera spp.on the muddy sediments and Sonneratia alba again

in the coraligenous sediments with the Excoecaria agallocha. Where waters are

Page 53 of 68

56

stagnant or where tidal inundation is less regular Rhizophora spp. and Avicennia

marina grow on muddy sediments in conjunction with Aegiceras corniculatum, Acanthus

ilicifolius, and Lumnitzera racemosa. On dark argilous more or less flooded soils which

are rarely inundated by sea water, Lumnitzera racemosa, Heritiera litoralis and

Acanthus ilicifolius are more common.In hinterland regions, in the brackish waters

where inundation occurs only in the rainy period, or along the margins of the streams

on dark argilous more or less muddy substrates Avicennia marina, Achrosticum

aureum, Xylocarpus Mekongensis, Corypha utan, Pandanus odoratissimus, Cycas

circinalis, Dolichandrone spathacea and Melaleuca leucadendra may occur. The latter

two species belong primarily to the landward margins and secondary forests of the low

coastal regions. Landward margins of the mangroves zones, not abutting with beaches,

generally contain salt flats being either bare muds or with samphire vegetation.

5.2.4 Current Uses and Threats

The illegal harvest and loss of mangroves remains a critical coastal management issue

in Timor Leste, with previous estimates indicating that total mangrove cover has being

reduced by approximately 30% between 1940-2000, from 9,000ha to just 3,035ha (FAO

2003). The current project indicates further significant losses of mangrove habitat, with

just 1,802 ha recorded in 2008, ie. ~40% loss between 2000-2008, or disturbingly, an

approximate 80% loss of total mangrove habitat in Timor Leste, since 1940.

Mangrove trees are harvested for timber and fuel wood and in some instances

hinterland mangroves have been removed for brackish water shrimp and/or fish ponds.

Harvesting for fuel wood is particularly intense in the Metinaro region (Alongi & De

Carvalho 2008), which has the largest, best developed mangrove stands in Timor Leste,

stretching over 750ha. This stand is currently under threat from uncontrolled

exploitation from a camp of Internally Displaced Persons (IDPs) that is a temporary

home to over 6,000 people (MTRC Information Centre 2006). Mangroves are exploited

for firewood and food stuffs such as bivalves, snails and fish species (Arafura and Timor

Seas Experts Forum 2006).

Page 54 of 68

57

5.3 Marine Habitats

Significantly, Timor-Leste’s north coast sits on the edge of an underwater precipice, the

Wetar Strait, a marine trench approximately 3 km deep, which provides a major corridor

for pelagic and migrating marine megafauna (ie. whales, dolphins, turtles, dugongs,

pelagic fish). In these ‘deep yet nearshore’ environments, deep localised upwellings

enrich Timor Leste’s nearshore coral reefs, and also, bring both predators and prey to

within reach of the coast. While the status of Timor-Leste’s coral reefs were predicted to

be ‘promising’ by Hodgson (1999) due to limited commercial exploitation, recent reef

fish surveys by Deutsch (2003), Wong and Chou (2004) and Dutra and Taboada (2006),

suggest otherwise. Apart from that, little or no research information on the extent or

status of the coral reefs of Timor Leste is available.

The current project has confirmed that while shallow, nearshore coral reef habitat in

Timor Leste is limited (~2000ha), with little lagoonal reef flat development (~458ha),

there is little evidence of habitat damage (ie. dynamite blasting, coral bleaching).

Coral-dominated communities tended to dominate the fore reef and escarpment,

particularly where the reef flat was narrow, with the most extensive nearshore coral

habitats occurring in the Dili to Metinaro-Manututo region (ie. 40km coastline east of

Dili) and in the Com-Jaco Island region. Significantly, a further 1266 ha of mixed coral-

seagrass and open reef flat was identified and also, 62,708ha of potential coral habitat

was identified in deeper waters. The latter underscores the need for more detailed

mapping of deeper water coral habitats. Similarly, the nearshore seagrass meadows

are very limited in extent (~2000ha), due largely to the absence of estuarine

environments and reef flats.

The very limited extent of coral reef, seagrass and mangrove habitats on the north

coast of Timor Leste, impose strong limits on available marine resources and levels of

harvest (particularly reef fisheries, mangroves) and in the light of increasing human

resource use, underscore the urgent need for precautionary and effective conservation

management.

Page 55 of 68

58

6 CAPACITY-BUILDING & ENGAGEMENT The project was well-supported by the project partners, the Ministry of Agriculture and

Fisheries (MAF) – with a wide range of training outcomes for MAF staff. The project

focussed on delivering the following training in coral reef monitoring, coastal marine

habitat mapping and surveys, and marine taxonomy.

6.1 International ‘Reef Check’ Training

Training in marine species identification and also, coral reef monitoring was considered

a major training priority for MAF staff. Training in survey, marine species identification

and ongoing monitoring of coral reef habitats (including fish, invertebrates) was

undertaken using the international, ‘ReefCheck’ protocol (www.reefcheck.org).

Under the TL ‘ReefCheck’ Project, 2 trainers (Shane Penny, Leo Dutra) were trained to

deliver ‘ReefCheck’ training to MAF staff. The first training session for MAF staff was

undertaken on 8-12th October 2007 in Dili. Field training sites (K41, Dollar Beach) were

chosen by the ReefCheck trainer, Shane Penny (with the assistance of Wayne Lovell,

FreeFlow Diving), following a 2-day, pre-training and logistics visit (11-13th September

2007). ‘ReefCheck’ training was undertaken using both, the SCUBA-based protocol

and the snorkel-based protocol – and with the training delivered both, in English and in

Indonesian – using Indonesian training material and products. A total of 12 MAF staff

were trained – with 3 certified as Eco-Divers (see Appendix 6).

Since this time, NRETAS has continued to regularly monitor/support MAF staff with

regular visits. MAF staff analysed and uploaded results onto the international,

‘ReefCheck’ database, with assistance from NRETAS.

6.2 Coastal & Marine Habitat Surveys

The coastal survey was supported by MAF staff (Fisheries, Protected Areas), with 2

staff assisting with the mangrove-saltmarsh surveys near Manututo and a further 2 MAF

staff (ALGIS, Forestry) assisting with the coastal mapping at the other locations.

Knowledge was shared on plant identification and vegetation structural description, as

well as GPS data collection techniques. Similarly, the marine survey had strong Page 56 of 68

59

engagement and participation by MAF. A total of 9 MAF staff (Fisheries, Forestry)

participated in the nearshore marine surveys, in addition to assisting with the

organisation of the field logistics (ie. liaison with local communities, boat hire,

accommodation, etc.). Prior to the marine field work, NRETAS ran a ½ day workshop

for MAF participants on the use of GPS, depth sounders and remote underwater

videos, and the benthic habitat classification.

6.3 Marine Taxonomy

Two MAF staff (Raphael Pereira, Fernando de Silva) received training in polychaete

taxonomy, as part of a 5-day, Marine Polychaete Taxonomy Workshop (1-5 October

2007) being held in Darwin, by the NT Museum & Art Gallery (NRETAS). Marine

polychaetes are a ubiquitous and dominant faunal group in tropical marine

environments, particularly in soft sediments. They are also a key indicator group for

assessing the impacts of marine pollution.

6.4 Regional Engagement

The Timor Leste coastal-marine habitat mapping program was presented at the recent

CDU-sponsored, regional GIS workshop in Kupang [‘GIS Applications for Sustainable

Development & Good Governance in Eastern Indonesia & Timor Leste’ (5-6 May,

Kupang)] and received very positive feedback. Significantly, government and research

participants at the workshop expressed considerable interest in the need to extend this

integrated research and training program to West Timor (and Flores).

7 EMERGING ISSUES & PRIORITIES

In consultation with key MAF staff (and other relevant GoTL agencies), the following

emerging issues and priorities for further work have been identified:

1. Following successful mapping and characterisation of the coastal and nearshore marine environments of the north and eastern coasts, further work is required on mapping the coastal and marine habitats of the south coast, and Oecussi, to complete a full mapping coverage at the national level.

Page 57 of 68

60

In identifying GoTL priorities for further coastal-marine habitat mapping, MAF have highlighted the need to address priorities for MPA network planning in Timor Leste - particularly, within the context of the broader, MPA network planning in the Lesser Sunda region being undertaken collaboratively with Indonesia (and The Nature Conservancy). To this end, MAF have identified the following priorities for further mapping: (1) west of Dili, including proposed sites for MPAs (high priority); (2) Oecussi (low priority); (3) south coast (low priority). Coastal (terrestrial) mapping and marine mapping were both identified as of equal importance.

2. The current marine and coastal habitat mapping and field survey program provides essential information to inform broad-scale conservation (ie. MPA) and regional coastal-marine resource planning and development (ie. fisheries, aquaculture, ecotourism). Finer-scale mapping however is required for local-scale, site-based management and sustainable resource use (ie. MPA zoning, developments) and in the short to medium term, is essential for future planning, sustainable use and management of the Timor-Leste coastal zone. Specifically, fine-scale mapping is required to assist the proposed zoning of the Nino Konis Santana Marine Park.

In identifying GoTL mapping priorities, finer-scale mapping for the Nino Konis Santana Marine Park was identified as the highest mapping priority, building on mapping being undertaken by ALGIS. Areas need to be prioritised for finer-scale mapping and a common classification scheme developed for the Timor Leste region. Capacity building was identified as high priority as it would deliver skills and knowledge for MAF (Fisheries/ALGIS) staff to specifically undertake mapping and spatial data analysis at both, the fine and broad-scale.

3. The current project has also highlighted the considerable opportunities (and

government interest) in extending the current coastal habitat mapping and training program to West Timor (and Flores).

MAF have identified capacity-building (training and equipment) in coastal-marine habitat mapping between Timor Leste, Australia (Northern Territory) and Indonesia (Nusa Tengarra Timor), as a high priority. Further, the need for a common classification scheme (and scales of information) and methods, across the region has also been identified as a key priority. This could be implemented through a regional, coastal mapping workshop, with participants from MAF (Fisheries/ALGIS), Nusa Tengarra Timor and the Northern Territory. This workshop would define terms of reference between the 3 countries on classification schemes and priorities and also, develop a plan of action. MAF also identified the need to train 2-3 staff in marine and coastal GIS and remote sensing techniques, in partnership with CDU.

4. Database training and knowledge management (and also, awareness of Intellectual

Property) is a priority area for training of MAF (and other government agency staff). With knowledge capture and the building of GIS and knowledge systems, training is required by MAF staff in database management, data standards, Intellectual Property and formal data agreements. MAF have also identified database training and knowledge management as a priority area for capacity-building and training of MAF and GoTL staff.

Page 58 of 68

61

5. MAF recognise the importance of Intellectual Property and knowledge management and recommend that all requests for data/information arising from work currently undertaken by MAF (and the Government of Timor Leste) (including this partnership program) should be subject to a formal data agreement.

6. Under the CTI National Plan of Action for Timor Leste, the Wetar Strait is priority

‘seascape’ requiring deep water habitat mapping. MAF, however, do not consider this a mapping priority, given the current priorities and lack of capacity and resources within MAF.

7. Given the success of the ‘ReefCheck’ program in Timor Leste, MAF have identified

the need to support and expand the existing MAF ‘ReefCheck’ survey program and team in Timor Leste. This includes establishing regular monitoring, 2-4 times year per site and increasing the number of sites across the northern coast. It was recommended that: (a) the reef survey program use ‘ReefCheck’ and also, other survey methods; (b) include impacted sites; and (c) establish regular monitoring sites within the Nino Konis Santana Marine Park (both, inside and outside the park).

8. MAF have identified the need to build on the TL ‘ReefCheck’ program (and results)

and develop a school-based, marine education and public awareness program, to highlight the status, condition and threats facing coral reefs in Timor Leste. In the Com-Tutuala region, this education campaign could be a component of a broader, Marine Park community outreach and education program.

8 REFERENCES Alongi DM & NA De Carvalho (2008) The effect of small-scale logging on stand characteristics

and soil biogeochemistry in mangrove forests in Timor Leste. Forest Ecology and Management

255: 1359-1366.

Arroyo LA, Healey SP, Cohen WB, Cocero D, Manzanera JA (2006) Using object-oriented

classification and high-resolution imagery to map fuel types in a Mediterranean region. Journal

of Geophysical Research-Biogeosciences 111

Audley-Charles, M.G., 2004. Ocean trench blocked and obliterated by Banda forearc collision

with Australian proximal continental slope. Tectonophysics 389, 65–79.

Barnett, J., Dessai, S., & Jones, R. (2003) Climate change in Timor Leste: Science, impacts,

policy and planning – Briefing to government, civil society, and donors, República Democrática

de Timor-Leste. The University of Melbourne & CSIRO. 40 pp.

Beech T, Dowd M, Field C, Hatcher B, Andrefouet S (2008) A stochastic approach to marine

reserve design: Incorporating data uncertainty. Ecological Informatics 3:321-333

Benz UC, Hofmann P, Willhauck G, Lingenfelder I, Heynen M (2004) Multi-resolution, object-

oriented fuzzy analysis of remote sensing data for GIS-ready information. Isprs Journal of

Photogrammetry and Remote Sensing 58:239-258

Page 59 of 68

62

Bock M, Xofis P, Mitchley J, Rossner G, Wissen W (2005) Object-oriented methods for habitat

mapping at multiple scales – Case studies from Northern Germany and Wye Downs, UK.

Journal for Nature Conservation 13:75-89

Bouma GA & HT Kobryn (2004) Change in vegetation cover in East Timor , 1989-1999. Natural

Resources Forum 28: 1-12.

de Carvalho, D. d. A., et al. (2007). Potensi Perikanan dan Potret Perempuan Zona Litoral

Kawasan Timur (potential of fishery resources and a portrait of women in the coastal zone of

eastern districts). Dili, Haburas Foundation.

Definiens (2006) Defiiniens Professional 5 User Guide. Definiens AG, Munchen, Germany

Deutsch C (2003). Coral Reefs in East Timor: An Investigation of the Impacts of Fishing on

Reef Fish Abundance. EM 531 Project Report, Unpublished report.

Diaz R, Solan M, Valente RM ( 2004) A review of approaches for classifying benthic habitats

and evaluating habitat quality. Journal of Environmental Management 73:165–181.

Dutra LXC & MB Taboada (2006) Preliminary Survey of the coral reefs from Beloi (Atauro

Island, Timor-Leste): alternatives for sustainable use. Ministry of Agriculture, Forestry and

Fisheries - National Directorate of Fisheries and Aquaculture, Dili, 9p.

ESRI (1996) Automation of Map Generalization: The Cutting-Edge Technology., vol. 2009.

ESRI

FAO (2003) Status and trends in mangrove area extent worldwide. Forest Resources

Assessment Working Paper No. 63. Rome, Forest Resources Division, FAO Rome. FAO (2007) The Worlds Mangroves 1980-2005. FAO Forestry Paper 153, FAO Rome.

Hajek F (2006) Object-oriented classification of Ikonos satellite data for the identification of tree

species composition. Journal of Forest Science (Prague) 52:181-187

Hamson, G. (2004) The tectonic evolution of East Timor and the Banda Arc. B.Sc.(Hons)

Literature Review, University of Melbourne.

Keep, M., Barber, L., & Haig, D. (2009) Deformation of the Cablac Mountain Range, East Timor:

An overthrust stack derived from an Australian continental terrace. Journal of Asian Earth

Sciences, 35:150–166.

Lyzenga DR (1978) Passive remote sensing techniques for mapping water depth and bottom

features. Applied Optics 17:379-383

Lyzenga DR (1981) Remote sensing of bottom reßectance and water attenuation parameters in

shallow water using aircraft and Landsat data. International Journal of Remote Sensing 2:71-82

Malthus TJ, Mumby PJ (2003) Remote sensing of the coastal zone: an overview and priorities

for future research. International Journal of Remote Sensing 24:2805-2815

Mumby PJ, Edwards AJ (2002) Mapping marine environments with IKONOS imagery:

enhanced spatial resolution can deliver greater thematic accuracy. Remote Sensing of

Environment 82:248-257

Page 60 of 68

63

Mumby PJ, Green EP, Clark CD, Edwards AJ (1998) Digital analysis of multispectral airborne

imagery of coral reefs. Coral Reefs 17:59-69

Mumby PJ, Green EP, Edwards AJ, Clark CD (1997) Coral reef habitat-mapping: how much

detail can remote sensing provide? Marine Biology 130:193-202

Mumby PJ, Harborne AR (1999) Development of a systematic classification scheme of marine

habitats to facilitate regional management and mapping of Caribbean coral reefs. Biological

Conservation 88:155-163

Mumby PJ, Skirving W, Strong AE, Hardy JT, LeDrew EF, Hochberg EJ, Stumpf RP, David LT

(2004) Remote sensing of coral reefs and their physical environment. Marine Pollution Bulletin

48:219-228

National Statistics Directorate (2006). Timor-Leste Census of Population and Housing 2004

Atlas. Dili.

Phinn SR, Dekker AG, Brando VE, Roelfsema CM (2005) Mapping water quality and substrate

cover in optically complex coastal and reef waters: an integrated approach. Marine Pollution

Bulletin 51:459-469

Sanderson PG (2001) The application of satellite remote sensing to coastal management in

Singapore. Ambio 30:43-48

RDTL and CDU (2006). The Timor-Leste Coastal/Marine Habitat Mapping for Tourism and

Fisheries Development Project, RDTL, Ministry of Agriculture, Forestry and Fisheries (MAFF),

Charles Darwin University (CDU).

Sandlund, O. T., Bryceson, I., de Carvalho, D., Rio, N., da Silva, J., & Silva, M.I. (2001).

Assessing Environmental Needs and Priorities in East Timor. Final Report. NINANIKU, UNDP,

Dili. Norwegian Institute for Nature Research, Trondheim.

UNDP and RDTL (2006). Environment & Natural Resource Management. Dili, Environment and

Natural Resource Management Unit, UNDP Timor-Leste.

UNDP and RDTL (2007). Capacity Building in and Mainstreaming of Sustainable Land

Management - SLM Project Factsheet. Dili, United Nations Development Programme, RDTL.

WB and ADB (2007). Economic and Social Development Brief, World Bank Group, Asian

Development Bank.

Wong LS. & LM Chou. (2004). Status of Coral Reefs on North-east Atauro Island, Timor-Leste,

based on surveys conducted in November 2004. REST Technical Report No. 7, Marine Biology

Laboratory, Department of Biological Sciences, National University of Singapore. 31p.

Page 61 of 68

64

APPENDICES

Appendix 1: Timor Leste inshore marine habitat field survey classification system.

Page 62 of 68

65

Appendix 2: Timor Leste marine habitat mapping field data survey sheet (and MS

Access database).

Page 63 of 68

66

Appendix 3: Timor Leste coastal vegetation field survey data collection sheet. Preliminary Vegetation Ground Truthing Sites: Timor-Leste Habitat mapping

Observers:

Site No:

Area: Manatuto, Metinaro, Baucau, Como, Jaco Date: 12 13 14 15 16 /2007

Preliminary

classification (habitat unit)

Georef: GZ 51

52 Wpt: Easting: North:

STRATA U1 M1 G1

% cover Av Ht (range) Growth Form Crown density

Dominant sps

% cover: Upper and Mid generally canopy cover: Ground generally percentage cover.(heights measured, cover an estimate)

Page 64 of 68

67

Appendix 4: Terrestrial plant species recorded in the Timor Leste coastal field

surveys.

Acacia leucaphloia Ipomea pes-caprae Acanthus ilicifolius Jatropha gossypiifolia Aegiceras corniculatum Lumnitzera racemosa Agele marmelos Ocimum tenuiflorum Avicennia marina Pemphis acidula Azima samentosa Pluchea indica Borassus flabellifer Premna serratifoliaCalotropis gigantea Prosopis pallida Ceriops australis Pterocaulon sphacelatumChromolaena odorata Rhizophora apiculata Cordea subcordata Rhizophora stylosa Corypha utan Schleichera oleosa Dichanthium annulatum Solanum spp Dichrostachys cinerea subsp malesiana Sonneratia alba Enneapogon polyphyllus Sonneratia caseolaris Eucalyptus alba Strychnos lucida Excoecaria agallocha Tamarindis indica Ficus spp Thespesia populneoides Fimbrystylis spp Vitex rotundifola Harrisonia brownii Vitex trifoliata Hyptis sauvolens Zyzyphus mauritiana

Page 65 of 68

68

Appendix 5: Comparison of Timor Leste field survey coastal vegetation groupings with broad mapping classes.

MapClass

Structure Groups Floristic Group

Upper Strata Species Obs

CF closed palmland 5 Borassus flabellifer-Schleichera oleosa-Agele marmelos 1 open forest 9 Corypha utan-Tamarindis indica,Tamarindis indica-

Corypha utan 1

CFM closed forest 1,22,222

Avicennia marina-Lumnitzera racemosa-Thespesia populneoide, Sonneratia alba-Rhizophora apiculata, Sonneratia alba

3

emergent- low open woodland

3,19 Avicennia marina, Rhizophora apiculata-Sonneratia alba-Aegiceras corniculatum

2

low closed forest 7,21 Ceriops sp-Avicennia marina, Rhizophora stylosa-Avicennia marina

4

low open forest 3,15,19,21,222

Avicennia marina,Lumnitzera racemosa, Rhizophora apiculata-Sonneratia alba-Aegiceras corniculatum, Rhizophora stylosa-Avicennia marina, Sonneratia alba

6

low woodland 1,3,7,222

Lumnitzera racemosa-Avicennia marina, Avicennia marina, Ceriops sp-Avicennia marina, Sonneratia alba

8

mid sparse shrubland

29 Lumnitzera racemosa-Jatropha gossypiifolia-Azima sarmentosa

1

nil upper 0 none 2 tall closed forest 20 Rhizophora mucronata 1 open forest 2,8,19,2

0,21,22,222

Avicennia marina-Sonneratia alba, Coconut palm-Sonneratia alba, Rhizophora apiculata, Rhizophora mucronata, Rhizophora stylosa, Sonneratia alba-Rhizophora apiculata, Sonneratia alba

11

woodland 10,13 Tamarindis indica, Excoecaria agallocha (BA-8)-Avicennia marina-Lumnitzera racemosa

2

DW nil upper 7 Ceriops decandra (shrubland) 1 OWS

emergent- low open woodland

16,18,24,26,27

P.weed-Corypha utan,Thespesia populneoides-Tamarindis indica,Zyzyphus mauritiana

6

low open forest 15,16 Lumnitzera racemosa, P.weed-Acacia sp-Zyzyphus mauritiana

2

low open palmland 9 Corypha utan-Tamarindis indica 1 low sparse

shrubland 18 Pluchea indica 1

low woodland 16 Species P 1 nil upper 0,27 none, Calotropis gigantea (shrub) 3 open forest 11 Eucalyptus alba-Borassus flabellifer 1 open palmland 5 Borassus flabellifer-Agele marmelos 1 woodland 10,12 Tamarindis indica, Eucalyptus alba 2 RFSL closed forest 21 Rhizophora stylosa-Sonneratia alba 1 nil upper 0 none 1 RP nil upper 26 grass (rice) 1 SF emergent- low

open woodland 27 Zyzyphus mauritiana-Cordea subcordata-Ficus 1

low open forest 17 Pemphis acidula 1 nil upper 0,7 none, Ceriops australis (shrubland) 6 W emergent- low

open woodland 10,27 Tamarindis indica-Acacia leucophloia-Zyzyphus

mauritiana,Zyzyphus mauritiana-Premna serratifolia-Schleichera oleosa

2

tall sparse shrubland

16 Prosopis pallida 1

low open forest 4 Borassus flabellifer-Acacia leucaphloia 1 low open palmland 5,99 Borassus flabellifer-Schleichera oleosa, Borassus

flabellifer-Zyzyphus mauritiana 4

low palmland 5 Borassus flabellifer-Agele marmelos 2 open forest 9 Tamarindis indica-Corypha utan 1

Page 66 of 68

69

Appendix 6: ‘Reef Check’ Training and Surveys in Timor Leste. Reef Check timeline of events 2007, October 9th –10th Ecodiver and snorkeller training, Dili & K41 2008, August 1st 1st all Timor-Leste Reefcheck survey at K41 2008, November 25th Reefcheck survey at Valu beach 2009, June 8th Reefcheck survey at K41

Reef Check Ecodivers and snorkellers

Ecodiver Carlos de Jesus Ministério da Agricultura e Pescas Ecodiver Celestino Barreto da Cunha Ministério da Agricultura e Pescas Ecodiver Anselmo Lopes Amaral Ministério da Agricultura e Pescas

Snorkeller Constancio Santos Silva Ministério da Agricultura e Pescas Snorkeller Jose Monteiro Ministério da Agricultura e Pescas Snorkeller Bendito Trindade Ministério da Agricultura e Pescas Snorkeller Lucas Fernandes Ministério da Agricultura e Pescas Snorkeller Narciso Almeida de Carvalho Ministério da Agricultura e Pescas Snorkeller Jose Nunes Ministério da Agricultura e Pescas

Reef Check survey sites K41 S 8o 28’ 30” E 125 o 53’

00” Valu Beach S 8o 24’ 40” E 127 o 18’

04”

Page 67 of 68

Appendix 6: ‘Reef Check’ Training and Surveys in Timor Leste. Reef Check timeline of events 2007, October 9th –10th Ecodiver and snorkeller training, Dili & K41 2008, August 1st 1st all Timor-Leste Reefcheck survey at K41 2008, November 25th Reefcheck survey at Valu beach 2009, June 8th Reefcheck survey at K41

Reef Check Ecodivers and snorkellers

Ecodiver Carlos de Jesus Ministério da Agricultura e Pescas Ecodiver Celestino Barreto da Cunha Ministério da Agricultura e Pescas Ecodiver Anselmo Lopes Amaral Ministério da Agricultura e Pescas

Snorkeller Constancio Santos Silva Ministério da Agricultura e Pescas Snorkeller Jose Monteiro Ministério da Agricultura e Pescas Snorkeller Bendito Trindade Ministério da Agricultura e Pescas Snorkeller Lucas Fernandes Ministério da Agricultura e Pescas Snorkeller Narciso Almeida de Carvalho Ministério da Agricultura e Pescas Snorkeller Jose Nunes Ministério da Agricultura e Pescas

Reef Check survey sites K41 S 8o 28’ 30” E 125 o 53’

00” Valu Beach S 8o 24’ 40” E 127 o 18’

04”

Page 67 of 68

70

(f(e

(d(c

(b(a

Page 68 of 68