PERSGA SAP. Component 5: A Regional Network of Marine Protected Areas.

PERSGA SAP. Component 5: A Regional Network of Marine Protected Areas. Survey of the proposed Marine Protected Area at Iles des Sept Frères and Ras Siyyan, Djibouti Draft Report October 2002 Nola Barker (International surveyor, UK) N. Galal (Regional surveyor, Egypt) Rebecca Klaus (Biotope and GIS/Remote Sensing Consultant, UK) 1

Executive Summary The Red Sea and Gulf of Aden contain some of the world s most diverse and biogeographically complex tropical marine habitats and communities, with the highest levels of endemism of any oceaniuc body of water. These factors make the Red Sea and Gulf of Aden globally important for marine biodiversity conservation. The primary goal of the Regional Organisation for the Conservation of the Environment of the Red Sea and Gulf of Aden (PERSGA) is the conservation of the environment, and consequently conservation of the biological diversity, of this unique region. The PERSGA Strategic Action Programme (SAP) supports and facilitates that goal (PERSGA, 1999). The SAP includes several components, one of which (Component 5) is Development of a Regional Network of Marine Protected Areas. Component 5 includes the development of a Regional MAster Plan for Marine Protected Areas (MPAs), and the development of a network of MPAs throughout the region (PERSGA 2001). Four MPAs of regional or global importance in the Red Sea and Gulf of Aden have been proposed to complement the already existing MPAs of the region. The survey, design and implementation of those four MPAs are to be carried out as part of Component 5 of the SAP and the Iles des Sept Frères and Ras Siyyan, Djibouti, is one of these proposed MPAs. Over the past forty years MPAs have been established throughout the world, and in the Red Sea and Gulf of Aden region, to conserve species, habitats and biodiversity, and more recently to support sustainable development. As such they act as tools to provide for sustainable use of natural resources, economic development and conservation. To inform the development of Site-specific Management Plans (Gladstone, 2001), ecological and resource use surveys of the proposed MPA were designed (Kemp, 2001) and carried out in 2001 in order to identify and map the distribution and diversity of marine and coastal habitats; to establish the current status of resources such fish and sessile invertebrates, and to assess the levels of human impact within the proposed MPA. The PERSGA MPA-SAP ecological field surveys utilise a combination of different survey methods, details for which can be found in Appendices 2-8. Site-specific data on resources including species abundance, distribution and diversity, together with human uses and impacts are key inputs to implementing plans and establishing an MPA (Gladstone, 2001). Rapid Ecological Assessment Methods (Price 2001, Kemp 2001) are increasingly used to provide this information readily on a broad scale, reducing the need for more time-consuming disciplinary research. As such its value has been increasingly recognised and methods developed to ensure valid comparability between regions (Price et al. 1988, Price 2001). Site specific surveys provide information only at survey sites, but information about the wider distribution and extent of marine habitats within the survey area is fundamental in management planning and zoning. The most cost-effective, efficient and well-established technique to obtain such information is to use satellite based remote sensing data and Geographical Information Systems (GIS) to map and quantify the areal extent of coastal and marine habitats. The PERSGA MPA-SAP surveys have employed the use of remote sensing data from the Landsat 7 Enhanced Thematic Mapper (ETM+) satellite data. The satellite data has been used to identify potential surveys sites prior to entering the field surveys and to create preliminary coastal and marine biotope maps using the ground truthed field survey data. During the PERSGA MPA-SAP field survey of the Islands of Sept Frères and Ras Siyyan a total of 178 geo-referenced locations were recorded. These consisted of 23 Rapid Ecological 2

Assessments, 10 Detailed Site Descriptions including 6 detailed Substrate and 6 detailed Fish Survey Transects, 69 Quick Site Surveys and 34 miscellaneous terrestrial measurements for use as ground control points (GCP) to georeference the satellite data. The field survey data have been entered into a GIS database and used to create point maps illustrating the distribution and status of resources within the proposed MPA. The Landsat 7ETM+ satellite image was classified using the ground truthed survey data, and preliminary marine and coastal biotope maps of the proposed MPAs were created. The proposed MPA at Sept Frères and Ras Siyyan covers a total area of 400km 2, including 107 km 2 of land (including the islands) and 293 km 2 of sea. The proposed MPA will include a total of 1.12km 2 of mangrove, 3km 2 of seagrass habitats, and 10km 2 of coral dominated habitats. The biotope maps created with the data gathered during this survey allow for the initial classification of land areas, however confirmation of this classification requires acquiring a second independent dataset. The consultants propose that such a dataset would be valuable information to be gathered at a later date. The recent survey revealed that most of the coral communities were in good condition, and that impacted sites were limited to areas experiencing human impacts, particularly at beaches and other popular landing and anchor sites at the islands. Fish community structure around the islands was also healthy, indicative of the low levels of fishing pressure. The exception to this was the low levels of sharks whose populations were depleted over the last few years by illegal fishing. Of particular concern was the abundance and distribution of Crown of Thorns starfish Acanthaster planci. Their presence in high numbers highlights the potential, if not an already apparent outbreak albeit requiring further research in order that it may be quantified. Previous research at Grand Ile and further south, at Iles Moucha report low numbers of A. planci, with the greatest number found in the south (PERSGA 1998). Birds were abundant throughout the survey area. Important areas for birds included all the islands of the Ile des Sept Frères and mangrove areas of Ras Siyyan (Fishpool & Evans, 2001). Human impacts were evident at the islands and on the mainland. Impacts were concentrated in certain locations, but away from these localised sites human impacts at present are low. Impacts observed were a direct result of human activity. This included tourist pressures of which spearfishing was one of the most detrimental in terms of potentially depleting key species, litter waste by tourists, fishers and other users including the military whose discarded weaponry was found on the beaches and underwater. Anchor damage was also evident in sheltered areas around the islands and along the coastline at Ras Siyyan. Data collected during the survey represent point samples over a wide area. These data may comprise baseline data for monitoring programmes to identify trends and patterns over time. Wide geographical coverage, variation and lack of repetition of assessments at sites due to time and various other constraints mean that data gathered in this survey provides an indication only of what species distribution and human impacts were present. Comparison between the islands with respect to coral cover for example would not be appropriate as generally only one transect was performed at each island. For this reason much larger sample sizes at each site will ultimately be required for long-term monitoring, however the information gathered so far allows the mapping of what sites are of significant importance to various species and where the main human impacts were. Finally, recommendations for further research are suggested. 3

Acknowledgements The authors wish to express their thanks and gratitude to the following for their help and assistance with the preparation and execution of this survey: Mr. Nasser Djama, PERSGA/SAP National Programme Co-ordinator, Djibouti; Mr. Houssein Rirach Robleh, Survey Coordinator, Direction de l Environnement (Environmental Management Department), Djibouti; Mr. Abdoulrazak Daoud, Commissaire, Obock; Mr. Abdoulkader Oudoum Abdallah, trainee, Direction de l'environnement cadre a l'amenagement du Territoire (Environmental Management and Regional Development), Djibouti; Mr. Ibrahim Omar Ali, trainee, Dolphin Excursions; Mr. Bruno Pardigon, General Manager, Dolphin Excursions (Dive tourism operators), Djibouti and Mr. Alex Hill and Ms Alice Hill, Dive tourism operators and PADI diving instructors affiliated to Dolphin Excursions. The success of the survey was also due to the enthusiasm, participation and professional assistance provided by Mr. Ibrahim Omar Ali, Captain Amin, Mr. Ali Z and Mr. Julien of Dolphin Excursions. Our thanks also go to the fishers and residents of Ras Siyyan and Obock for their participation and informative discussions. 4

Table of Contents 1 INTRODUCTION... 9 1.1 THE PHYSICAL ENVIRONMENT... 9 1.1.1 The Djibouti coast... 9 1.2 THE BIOLOGICAL ENVIRONMENT... 10 1.3 THE SOCIO-ECONOMIC ENVIRONMENT... 10 1.4 THE SURVEY AREA... 11 2 OBJECTIVES OF THE SURVEY... 14 3 METHODS... 14 3.1 PREPARATION OF LANDSAT 7ETM+ SATELLITE IMAGE DATA... 14 3.2 FIELD SURVEY TIME FRAME... 14 3.3 SURVEY DATA COLLECTED... 15 3.4 TRAINING AND CAPACITY BUILDING... 15 4 RESULTS... 24 4.1 THE MARINE AND COASTAL HABITATS AND BIOTOPES OF SEPT FRÈRES AND RAS SIYYAN.... 24 4.1.1 Coastal Habitats and Biotopes... 24 4.1.2 Marine Habitats and Biotopes... 29 4.2 THE ECOLOGICAL ENVIRONMENT... 34 4.2.1 Coral communities... 34 4.2.2 Seagrasses... 39 4.2.3 Mangroves... 40 4.2.4 Fish, Crown of Thorns starfish, sea urchins, sea cucumbers, clams and gastropod communities... 43 4.2.5 Crown of Thorns Starfish... 48 4.2.6 Turtles... 52 4.2.7 Birds... 53 4.2.8 Cetaceans... 55 4.2.9 Dugong... 55 4.2.10 Shark... 55 4.2.11 Terrestrial invertebrates, mammals and reptiles... 55 4.2.12 Other oceanic megafauna... 57 4.3 REPORT ON THE INTERTIDAL ENVIRONMENT (See Appendix 16).... 57 4.3.1 Executive summary... 57 4.4 REPORT ON THE SOCIOECONOMIC ENVIRONMENT (See Appendix 16).... 58 4.4.1 Centres of population... 58 4.4.2 Size of population... 58 4.4.3 Occupations... 58 5

4.5 REPORT ON THE FISHERIES (See Appendix 16).... 58 4.6 HUMAN IMPACTS... 58 4.6.1 Construction... 58 4.6.2 Fishing/collecting... 59 4.6.3 Pollution... 59 4.6.4 Tourism... 62 4.7 NATIONAL CAPACITY BUILDING IN MPA SURVEY, DESIGN AND MANAGEMENT... 62 5 CONCLUSION... 63 5.1 BIOLOGICAL ENVIRONMENT... 63 5.2 HUMAN IMPACTS... 63 5.3 RECOMMENDATIONS FOR FURTHER SURVEY WORK... 63 6 REFERENCES... 65 APPENDIX 1: Relevant sections of the terms of reference....66 APPENDIX 2: Survey sheet 1 (Rapid Ecological Assessment).69 APPENDIX 3: Survey sheet 2 (Site Description)..71 APPENDIX 4: Survey sheet 3 (Subtidal Transect Sheet)... 72 APPENDIX 5: Survey sheet 4A (Reef Animal Belt Transect)...75 APPENDIX 6: Survey sheet 4B (Reef Fish Community Structure)..76 APPENDIX 7: Survey sheet 5 (Quick Site Surveys)......77 APPENDIX 8: Survey sheet 6 (Intertidal Site Description Sheet). 78 APPENDIX 9: Field Survey itinerary 79 APPENDIX 10: Survey Data Sheet Record 81 APPENDIX 11: Coral Transect Data.. 82 APPENDIX 12: Mapping Marine and Coastal Biotopes 86 APPENDIX 13: Summary of Satellite Image Header File Documentation for Djibouti...99 APPENDIX 14: Atmospheric Correction of Landsat 7ETM + Satellite Data.100 APPENDIX 15: MapInfo GIS Data Layers for Djibouti. 103 APPENDIX 16: Socioeconomic, Fisheries and Intertidal reports, Djibouti...104 APPENDIX 17: Report on the trainees accompanying the survey of the Iles des Sept Frères and Ras Siyyan proposed MPA, Djibouti...162 APPENDIX 18: Problems encountered during the ecological survey of the Iles des Sept Frères and Ras Siyyan proposed MPA, Djibouti...166 6

List of Figures: Figure 1. Map of Djibouti... 9 Figure 2: True colour composite of Landsat 7ETM satellite image of Sept Frères and Ras Siyyan (WRS166/051 13 th May 2000).... 13 Figure 3: Rapid sites surveyed, Djibouti 2002... 17 Figure 4: Quick look sites, Djibouti 2002.... 18 Figure 5: Unsupervised classification of marine areas in the Landsat 7ETM+ satellite image of Sept Frères and Ras Siyyan, Djibouti (WRS166/051 13 th May 2000).... 19 Figure 6 : Unsupervised classification of coastal areas in the Landsat 7ETM+ satellite image of Sept Frères and Ras Siyyan, Djibouti (WRS166/051 13 th May 2000)... 20 Figure 7: A subsection of the unsupervised classification of marine areas in the Landsat 7ETM+ satellite image of Ras Siyyan, (WRS166/051 13 th May 2000) created prior to the field survey.... 21 Figure 8: A subsection of the unsupervised classification of coastal areas in the Landsat 7ETM+ satellite image Ras Siyyan, (WRS166/051 13 th May 2000) created prior to the field survey... 22 Figure 9: The coastal habitats and biotopes of Sept Frères and Ras Siyyan Djibouti (Landsat 7ETM+ 166/051 13 th May 2000)... 25 Figure 10: The coastal habitats of Sept Frères and Ras Siyyan... 26 Figure 11: The marine biotopes of Sept Frères and Ras Siyyan Djibouti (Landsat 7ETM+ 166/051 13th May 2000).... 30 Figure 12: Pie chart showing the area of marine biotopes mapped as a proportion of the total area mapped (excluding deep water areas).... 33 Figure 13: Substrate Transect data, Djibouti 2002... 37 Figure 14: A map illustrating the distribution of coral habitats around Sept Frères and Ras Siyyan. Dense coral habitats (in red) indicate where the area is dominated by corals. Sparse coral habitats (in pink) indicate areas where corals were found in association with other more dominant groups such as macroalgae or turf algae.... 38 Figure 15: The distribution of seagrass biotopes (areas highlighted in purple) around Sept Frères and Ras Siyyan.:... 39 Figure 16: Mangroves at Rapid Sites, Djibouti 2002... 41 Figure 17: The distribution of mangroves (areas highlighted in green) around Ras Siyyan Djibouti as identified from the Landsat 7ETM satellite image.... 42 Figure 18: Abundance of fish and animals from fish transects, Djibouti. Other consists of various fish and other animals (clam and urchins)... 43 Figure 19: Relative abundance of different fish groups, Djibouti 2002.... 44 Figure 20: Relative fish abundance at Rapid Sites, Djibouti 2002... 45 Figure 21: Abundance and distribution of fish groups at six sites, Djibouti 2000... 46 Figure 22: Distribution and abundance of benthic fauna, Djibouti 2002.... 47 Figure 23: Distribution and abundance of A. planci (COTs) scars at Rapid Sites, Djibouti 2002... 50 Figure 24: Distribution and abundance of A. planci (COTs) scars at Quick Sites, Djibouti 2002... 51 Figure 25: Distribution and abundance of turtle pits at Rapid Sites, Djibouti 2002.... 52 Figure 26: Bird density and distribution from Rapid Site surveys, Djibouti 2002.... 54 7

Figure 27: Distribution and abundance of invertebrates at Rapid Sites, Djibouti 2002.... 56 Figure 28: Distribution and abundance of litter at Rapid Sites, Djibouti 2002.... 61 LIST OF TABLES: Table 1: Land area within the proposed MPA at Ras Siyyan and Sept Frères (from the Landsat 7ETM satellite data).... 24 Table 2: The marine biotopes mapped in the Landsat 7ETM+, the area estimates and the cluster groupings identified from cluster and SIMPER analysis are also included.... 31 Table 3: The areas of marine biotopes mapped and the % of the marine biotopes mapped as a proportion of the total area mapped (293.78 km 2 ) and the total area mapped excluding deep water classes (40.5 km 2 ).... 32 Table 4: The total area of coral habitats within the Ras Siyyan and Sept Frères proposed MPA (estimated from the marine biotope classification of Landsat 7ETM+ data).... 34 Table 5. Bird species identified... 53 Table 6. Species and distribution of dolphin seen in the Ras Siyyan and Iles des Sept Frères area, Djibouti June 2002.... 55 8

1 INTRODUCTION 1.1 THE PHYSICAL ENVIRONMENT 1.1.1 THE DJIBOUTI COAST Djibouti is located in Eastern Africa at the juncture of the Red Sea and the Gulf of Aden, with its borders adjoining Eritrea, Ethiopia and Somalia (Fig. 1). Djibouti covers an area of 23,200 km 2 and has a coastline of 370km. The climate is hot and semi-arid with mean air temperatures varying between 25ºC in winter and 35ºC in summer. Rainfall is between 50 and 215mm per year, averaging 130mm. Seawater salinity varies from 36.5ppt and 39ppt depending on water flow (Edwards, 1987), with mean surface water temperatures ranging from 25ºC and 29ºC. Survey Area Figure 1. Map of Djibouti The Bab al Mandab, 29km wide, links the Red Sea to the Indian Ocean. It s relative shallowness (100-130m deep) restricts the exchange of water between the two areas. The progressively decreasing width of the continental shelf in the Red Sea from a few km offshore in the north, to 9

about 100km wide in the southern Red Sea has led to a progressive change in the range of marine habitats present. Changes in water density are caused by variations in temperature, salinity and winds. During the summer monsoon (approximately June to September) winds blow from the west to east through the Gulf of Aden, blowing warm surface waters offshore. These warm waters are replaced by cool, nutrient rich water from the depths of the Gulf of Aden and Arabian Sea (Sheppard et al. 1992). This results in high pelagic primary productivity, measured as total pigment concentrations, extending from the Gulf of Aden northwards into the Red Sea to 19 N (Sheppard et al. 1992). In the winter monsoon (approximately October to May) winds blow from east to west, from the Gulf of Aden into the Red Sea. The movement of water as described, results in upwellings in the region of the Bab-al-Mandab providing the basis of the southern Red Sea fisheries. 1.2 THE BIOLOGICAL ENVIRONMENT The Bab-al-Mandab links the Red Sea to the Gulf of Aden and the rest of the Indo-Pacific. Due to its unique geographical position the area is of great biological importance (See Sheppard et al. 1992 for overview and Kemp and Benzoni, 2000) and much of the mainland and island coastlines are dominated by coral growth. A number of marine surveys of the area have been carried out including those of PERSGA (1998), Obura (1998) and Obura & Djama (2000). These provide preliminary description of many of the features of the area. Much shallow hard substrate surrounding the islands and the mainland of the survey area are dominated by coral communities. Some form well-developed coral reefs such as those found fringing the southern part of the mainland but many, particularly around the islands, are corals living on volcanic or basalt-like substrate, known as rocky (non-coral) reefs. Two significant areas of mangroves were located in the survey area, one at Ras Siyyan and the other at Khor Angar. Seagrass beds were found within the lagoon at Ras Siyyan but appeared sparce. Due to Djibouti s proximity to Arabia, it is also an important migration route for birds travelling between Europe and the African continent. For additional information on biodiversity and ecology of the area, see PERSGA, 2002; Kemp & Benzoni, 2000 and Ahmed et al., 2001. 1.3 THE SOCIO-ECONOMIC ENVIRONMENT The population of Djibouti is estimated at 557,000 (PERSGA, 2001). Two-thirds to threequarters of the population is located in the capital, Djibouti. The remainder of the non-nomadic population is split between two towns, Tadjoura and Obock and various scattered villages and settlements. The exact number of nomadic people is unknown. The population growth rate is estimated at 4.9% (PERSGA, 2001), with 43% of the population below the age of 14 years. Average life expectancy is 49 years (PERSGA, 2001). In addition to the local population is the presence of a non-african expatriate community, numbering between 10,000 and 15,000 people and immigrants from Yemen, Ethiopia and Somalia. The economy of Djibouti is largely based on service activities connected with the country s strategic location, proximity to Arabian oil fields and status as a free trade zone in northeast Africa. Djibouti is both a transit port for the region and an international trans-shipment and re-fuelling centre. In 2000, Gross Domestic Product (GDP) was estimated at DJF102,012 (US$574 million) with a growth rate of 2% (CIA, 2002). 75% of the active labour force is involved in agriculture, however scanty rainfall limits crop production to fruit and vegetables, and together with animal breeding and fisheries these activities contribute some 3% to GDP. 10

Smaller proportions of the labour force (11% and 14% respectively) are involved in industry and services respectively (1991 estimate CIA, 2002), contributing 22% and 75% to GDP (1998 est. CIA). Unemployment ranges between 40 to 50% (2000 estimate CIA, 2002). Tourism, and in particular marine tourism is an option available to Djibouti that would diversify the economy, making use of the country s natural resources. Djibouti s coral communities and other marine and coastal resources could, if sustainably managed, prove to be a valuable asset for tourismbased industries. 1.4 THE SURVEY AREA The proposed MPA encompasses the Iles des Sept Frères (Islands of the Seven Brothers) and the mainland coast from north of Ras Siyyan southwards to Khor Angar. The southern boundary of the proposed MPA lies approximately 100km north of Djibouti (Figs. 1 & 2). The area extends approximately 20km from north to south and 15km offshore, including approximately 30km of mainland coastline in addition to the coastlines of all six islands lying in the narrow strait of the Bab-al-Mandab. 11

Figure 2: True colour composite of Landsat 7ETM satellite image of Sept Frères and Ras Siyyan (WRS166/051 13 th May 2000). 13

2 OBJECTIVES OF THE SURVEY The objective of the survey was to carry out ecological and resource use surveys of the Proposed Marine Protected Area at Iles des Sept Frères and Ras Siyyan, Djibouti, including: 1. Provision of baseline information to guide the development of a Site Specific Master Plan for the proposed Iles des Sept Frères and Ras Siyyan MPA, that baseline information to include: Identification of the distribution of the principal habitats and species within the proposed MPA. Carrying out a preliminary assessment of the principal resource uses within the survey area. Carrying out a preliminary assessment of the probable threats to sustainable use and conservation within the survey area. Determining the extent of the area to be included within the MPA. To meet with stake-holders throughout the survey area, to inform them of the purpose and as far as possible of the preliminary findings of the surveys, and to give them the opportunity to provide input to the MPA project at the earliest stages. (See Kemp 2001 for Survey Design) 2. Building national capacity in Marine Protected Area survey, design and management through on-the-job training. 3 METHODS To fulfil the objective of collecting baseline information as noted above, methods detailed in PERSGA Training Manual. Documents 1 to 5 and corresponding survey sheets were used (see Appendices 2-8). For additional details pertaining to the methods see Survey Design for Proposed Marine Protected Areas (Kemp, 2001). 3.1 PREPARATION OF LANDSAT 7ETM+ SATELLITE IMAGE DATA Prior to the start of the field phase, the Landsat 7ETM+ satellite image of Sept Frères and Ras Siyyan (WRS 166/051, 13 th May 2000) (Figure 2) was atmospherically and geometrically corrected. Unsupervised classification of marine (ETM1-3 Figure 3) and coastal (ETM3-5 Figure 4) areas within the proposed MPA were prepared. The maps were subsetted (Figure 5 and Figure 6) printed and laminated for use during the field surveys. Further details on the Landsat satellite image and the image processing methods are included in Appendix 11. 3.2 FIELD SURVEY TIME FRAME The time frame designated for the field surveys was from the 8 th to the 23 rd June 2002. Details of Field Itinerary are listed in Appendix 9. For the duration of the data collection in the field, the team was based on a live-aboard dive boat. A 7m long open fishing boat was used to get to all areas including the islands, the lagoon and points on the mainland. 14

3.3 SURVEY DATA COLLECTED Rapid ecological assessments (Survey sheet1, Appendix2) were undertaken at 23 sites (Fig.3), with detailed site descriptions (Survey sheet 2, Appendix 3) of 10 more sites throughout the MPA proposed area. Underwater surveys included substrate study transects (Survey sheet 3, Appendix 4) at six sites and reef animal belt transects (Survey sheet 4A, Appendix 5) at the same locations. Quick look surveys (Fig.4; Survey form 5, Appendix 7) were done by calculating the mean value from several samples each sample consisting of a 10x10m quadrat. Appendix 10 details the number and type of survey carried out in the study area. No surveys were conducted along the fore reef slope at Ras Siyyan due to poor weather conditions. In addition, three other survey teams undertook intertidal (Survey sheet 6, Appendix 8), fisheries and socio-economics survey work, individual reports for which have been completed (see Appendix 16 for full reports). Where available, executive summaries for the latter three surveys have been incorporated into the results section of this report. The results of the field survey from 173 survey sites around the proposed MPA were entered into Geographical Information System (MapInfo TM Professional v6.0). A list of the GIS data layers created is included in Appendix 15. Data from these surveys were used to identify the main coastal and marine habitat and biotopes. Figures 5 to 8 show the unsupervised classification of marine and coastal areas of the region. The biotopes were identified using multivariate analysis of the detailed and quick site data, using hierarchical agglomerative clustering (Wards Method) and SIMPER analysis. These data were used to perform a supervised classification of the Landsat 7ETM+ satellite data to identify and map the extent of the coastal and marine biotopes (details of which are included in Appendix 12). 3.4 TRAINING AND CAPACITY BUILDING In order to build national capacity in MPA survey, design and management, two Djiboutian nationals were nominated as trainees to accompany and take part in the survey throughout the field phase. Trainees were to rotate between each of the three survey teams (Fisheries, Socio- Economics and Ecological/sub-tidal teams) so as to gain experience from all aspects of the different survey methods and preparatory work involved in collecting baseline information prior to establishment of a proposed MPA. Interaction and teaching throughout the survey prior to, during and after daily fieldwork were also integral parts of the training exercise. Prior to the survey, all training documents and survey sheets were translated into French to facilitate training of any non English-speaking trainees. 15

Figure 3: Rapid sites surveyed, Djibouti 2002 17

Figure 4: Quick look sites, Djibouti 2002. 18

Figure 5: Unsupervised classification of marine areas in the Landsat 7ETM+ satellite image of Sept Frères and Ras Siyyan, Djibouti (WRS166/051 13 th May 2000). 19

Figure 6 : Unsupervised classification of coastal areas in the Landsat 7ETM+ satellite image of Sept Frères and Ras Siyyan, Djibouti (WRS166/051 13 th May 2000) 20

Figure 7: A subsection of the unsupervised classification of marine areas in the Landsat 7ETM+ satellite image of Ras Siyyan, (WRS166/051 13 th May 2000) created prior to the field survey. 21

Figure 8: A subsection of the unsupervised classification of coastal areas in the Landsat 7ETM+ satellite image Ras Siyyan, (WRS166/051 13 th May 2000) created prior to the field survey 22

4 RESULTS Section 4.1 presents an overview of the results of the supervised classification of marine and coastal biotopes created using the Landsat 7ETM+ satellite image. The detailed results of analysis of field survey data and satellite image processing methods used are included in Appendix 12. In Section 4.2 the results from the sites specific field surveys (Rapid, Detailed and Quick Site surveys) and the marine and coastal biotope classification are combined and discussed by major habitat types namely corals, seagrass, mangroves, then fish and other communities. These results are then followed by results of Rapid surveys on the socioeconomic environment and human impacts. 4.1 THE MARINE AND COASTAL HABITATS AND BIOTOPES OF SEPT FRÈRES AND RAS SIYYAN. The proposed MPA at Sept Frères and Ras Siyyan, Djibouti covers an area of 400km 2, which includes 107 km 2 of land and 293 km 2 of sea. The Landsat 7ETM+ satellite image has been analysed to create preliminary estimates on the distribution and extent of different marine and coastal biotopes within this area, which are described further below. Table 1: Land area within the proposed MPA at Ras Siyyan and Sept Frères (from the Landsat 7ETM satellite data). Name Area (km 2 ) Djibouti mainland coast 105.998 Ile du Sud 0.042 Ile de l'est 0.222 Grande Ile 0.539 Ile du Bas 0.025 Boeing 0.094 Hamra 0.074 Siyyan Himar 0.006 Total Area Land 107 Total Area Marine 293 Total Area 400 4.1.1 COASTAL HABITATS AND BIOTOPES The coastal biotope map of Sept Frères and Ras Siyyan (Fig. 9) was created using the site descriptions from the Rapid Assessment surveys and from site specific photographic observations. Detailed subsections of the coastal biotope map are included in the Appendix 11 (Figures A14-A16). The islands of Sept Frères predominantly consist of eroded volcanic ash or tufa cones (Figure 10a and 10b) with the exception of Hamra which is formed from a red basalt similar to that at the headland of Ras Siyyan (Figure 10c). The coast around Ras Siyyan consists of a mixture of undercut fossil reef platform (Figure 10e and f) and sand interspersed with basalt rock outcrops. The mainland coast north of Ras Siyyan consists mainly of sand with a mixture of mud, sand and gravel sediments with areas of sparse halophytic vegetation (Figure 10g and h). 24

Rock: volcanic Figure 9: The coastal habitats and biotopes of Sept Frères and Ras Siyyan Djibouti (Landsat 7ETM+ 166/051 13 th May 2000) 25

Further south, along the northern edge of the peninsula leading to Ras Siyyan, is an extensive area of mangrove (Figure 10h and 10i). The mangroves are described in greater detail in section 4.2.3 Areas of sparse halophytic vegetation were also found. In the middle of the peninsula set behind (south) of the mangrove are two large areas of sabka (Figure 10h). The coast along the southern edge of the peninsula (east) is predominantly sand with rock outcrops. Further south along the mainland coast towards Khor Angar the coast is again dominated by mangrove. The mangrove fringes the edge of the Khor which is open directly into the sea. This area of mangrove extends further south beyond the boundaries of the MPA. Figure 10: The coastal habitats of Sept Frères and Ras Siyyan a. Eroded volcanic ash / cone (Sept Frères). The white marks on the cliffs and shore are areas of guano from the abundant roosting birds. b. Weathered beach rock c. Basalt rock at Ras Siyyan headland 26

d. Basalt gravel /scree (Ras Siyyan) e. Undercut fossil reef rock (NW Ras Siyyan). f. Fossil reef rock g. Gravel and Sand Plain (Ras Siyyan in distance) 27

h. Sand and Sabkha and Mangrove (in distance) (view south from Ras Siyyan) i. Mangrove (Avicennia marina) j. Sparse vegetation (Panikamp grass) on sand and gravel. 28

4.1.2 MARINE HABITATS AND BIOTOPES Field data from the 79 Quick and Detailed site surveys were analysed using hierarchical agglomerative clustering and SIMPER analysis to identify the different biotopes categories (Fig.11). A biotope is defined the physical habitat together with it associated assemblage of species. As no species information were collected during these surveys the use of the term biotope in this context is fairly loose and refers to the benthic structure and dominant cover types. The results of the cluster analysis and SIMPER (Appendix 11) revealed 16 broad marine biotopes. The categories of marine biotope identified from the cluster analysis of the ground truth data, together with additional descriptive data from the Rapid site surveys, were used to develop of a set of 51 spectral signatures (excluding deep water) to enable a supervised classification of the image. The signature set was then used to conduct a supervised classification of the Landsat 7ETM satellite data. A total of 21 marine biotope classes (Table 2) were mapped including 6 sand and sediment classes, 1 seagrass and 3 algal biotopes, 1 carbonate rock biotope and 8 coral biotopes with the remaining classes being deep or turbid water. The supervised classification of marine habitats and biotopes of Sept Frères and Ras Siyyan from the Landsat 7ETM+ satellite image is presented in Figure 11. Detailed subsections of the marine biotope maps are included in the Appendix 11 (Figures A20-A26). Table 3 lists the area covered by the different marine biotopes mapped and the proportion of the marine biotopes by the total area mapped and the total area excluding deepwater. The proposed MPA covers an area of 293 km 2 of marine habitats and over 86% of this area is in deep water greater which was not possible to map using Landsat satellite data. The coral and seagrass habitats are discussed in more detail below (Section 4.1.3 and 4.1.4). The sand habitats are also common covering over 6 km 2. Figure 12 shows the area of marine biotopes mapped as a proportion of the total area mapped (excluding deep water areas). 29

Figure 11: The marine biotopes of Sept Frères and Ras Siyyan Djibouti (Landsat 7ETM+ 166/051 13th May 2000). 30

Table 2: The marine biotopes mapped in the Landsat 7ETM+, the area estimates and the cluster groupings identified from cluster and SIMPER analysis are also included. Marine Biotope Classification Area (km 2 ) Deep water 253.29 NA Deep sand / rubble / rock (>10m) 10.75 14 Sand (5-10m) 4.66 7 Sand (2-5m) 7 Sand (1-2m) 7 Sand (<1m) 4.85 7 Sand and silt 2.35 7 Turbid water in Khor 0.13 NA Turf algae (15%) & sparse macroalgae (5%) & hard coral (3%) on sand and rock 2.21 11 Turf algae (2%) and sponge (2%) on sand and rubble 0.06 13 Macroalgae (45%) and sparse coral (5%) on sand and rubble 0.84 8 Sand and seagrass (15%) 3.20 15 Reef flat / carbonate platform 2.22 Sparse hard coral (Pocillopora) (10%) and macroalgae (5%) on sand and rubble 3.21 10 Dense hard coral (Acropora) (60%) on sand and rubble 1.40 9 Massive hard coral community (40%) on sand 0.99 2 Mixed hard coral community (20%) on sand and rubble 0.01 12 Fore reef slope 2.36 NA Massive hard (30%) and soft corals (10%) on rock and sand 0.17 6 Cluster Groups Mixed hard and soft coral community on rock 0.64 3,4,5,16 Mixed hard and soft coral community on rock, sand and rubble 0.45 1 Total Area 293.78 31

Table 3: The areas of marine biotopes mapped and the % of the marine biotopes mapped as a proportion of the total area mapped (293.78 km 2 ) and the total area mapped excluding deep water classes (40.5 km 2 ). Marine Biotope Classification Area (km 2 ) % of Total Area Mapped Deep water 253.29 86.21 % of Total Area (excl. deep water) Deep sand / rubble / rock (>10m) 10.75 3.66 26.54 Sand (<1m), Sand (1-2m), Sand (2-5m), Sand (5-10m) 9.51 3.24 23.48 Sand and silt 2.35 0.80 5.80 Turbid water in Khor 0.13 0.04 0.32 Turf algae (15%) & sparse macroalgae (5%) & hard coral (3%) on sand and rock 2.21 0.75 5.46 Turf algae (2%) and sponge (2%) on sand and rubble 0.06 0.02 0.15 Macroalgae (45%) and sparse coral (5%) on sand and rubble 0.84 0.29 2.07 Sand and seagrass (15%) 3.20 1.09 7.90 Reef flat / carbonate platform 2.22 0.76 5.48 Sparse hard coral (Pocillopora) (10%) and macroalgae (5%) on sand and rubble 3.21 1.09 7.93 Dense hard coral (Acropora) (60%) on sand and rubble 1.40 0.48 3.46 Massive hard coral community (40%) on sand 0.99 0.34 2.44 Mixed hard coral community (20%) on sand and rubble 0.01 0.00 0.02 Fore reef slope 2.36 0.80 5.83 Massive hard (30%) and soft corals (10%) on rock and sand 0.17 0.06 0.42 Mixed hard and soft coral community on rock 0.64 0.22 1.58 Mixed hard and soft coral community on rock, sand and rubble 0.45 0.15 1.11 32

Figure 12: Pie chart showing the area of marine biotopes mapped as a proportion of the total area mapped (excluding deep water areas). 33

4.2 THE ECOLOGICAL ENVIRONMENT 4.2.1 CORAL COMMUNITIES Figure 11 illustrates the distribution of coral biotopes found around Sept Frères and Ras Siyyan (see previously). A total of 8 broad coral biotopes were identified. Figure 14 illustrates the combined areas for all coral biotopes. The coral biotopes have been assigned as either dense coral habitat where coral is the dominant cover or sparse coral habitat where coral communities were found in association with other groups. The total area of coral habitats is 11.45 km 2, with a total of 10.05 km 2 in Sector 1 and 1.39 km 2 in Sector 2. Table 4: The total area of coral habitats within the Ras Siyyan and Sept Frères proposed MPA (estimated from the marine biotope classification of Landsat 7ETM+ data). Sector 1 Sector 2 Total Area Coral Habitats (km 2 ) Ras Siyyan (km 2 ) Sept Frères (km 2 ) Sparse coral communities 6.24 6.08 0.16 Dense coral 5.21 3.97 1.23 Total Coral Habitats 11.45 10.05 1.39 4.2.1.1 Coral Communities Sector 1 (Ras Siyyan) The mainland coast north and south of Ras Siyyan is bordered by a wide discontinuous fringing reef broken by deep water channels (to the north of Ras Siyyan), areas of mobile sediment (to the south of Ras Siyyan) and freshwater outflow / sediment, opposite Khor Angar as illustrated in Figure 11. The fringing reef north of Ras Siyyan was backed by a broad sedimented lagoon and wide area of reef flat. The lagoon was predominantly shallow (1-4m deep) with a sandy bottom that was frequently covered with a thin layer of mud supporting thin algal mats and scattered patches of macroalgae including Padina, Caulerpa, Sargassum and Turbinaria and seagrass beds (see below). There were sparse but extensive coral communities in the lagoon at Ras Siyyan but percentage cover was typically <5% dominated by Pocillopora sp. (Figure 11 and Figure 14). The fringing reef was broken by a deep water channel leading into the southern end of the lagoon just north of Ras Siyyan. The headland at Ras Siyyan was fringed by a narrow discontinuous reef with branching Acropora and Porites. A narrow fringing reef also extended within the lagoon along the edge of the deep water channel. Extending southwards from the headland at Ras Siyyan there is a more or less continuous and quite well developed fringing reef extending to Khor Angar. Hard coral communities are dominated by Acropora and Pocillopora. Coral cover ranged between 10% and 55% with a mean value of 25%. (See Appendix 11 for coral graphs). Approximately 1.5km offshore from Khor Angar, is a lozenge shaped area of ribbon reef (N12 25.835, E43 20.293) development orientated parallel to the shore (Fig 11). The top of the ribbon reef structure was at approximately 2-3m depth and was approximately 200m wide and 3.5km in length. This structure was separated from the mainland coast by an area of deep sand. The fringing reef then continued along the shore south of Khor Angar. 34

4.2.1.2 Coral communities Sector 2 (Islands of the Sept Frères) The islands of Sept Frères were fringed by well developed coral communities which did not generally form coral reefs but were typically found growing on non-biogenic rock which typically descend directly from the shore with little or no reef development (see Sheppard & Salm, 1988 for categories). The islands are exposed to significant wave action that has resulted in the development of a shallow groove and spur system. Three coral biotopes were common around the Sept Frères Massive hard (30%) and soft corals (10%) on rock and sand (0.17km 2 ), Mixed hard and soft coral community on rock (0.64km 2 ) and Mixed hard and soft coral community on rock, sand and rubble (0.45km 2 ). Weathered rock which occurred along the majority of the shoreline surrounding the islands was typically covered with a thin layer of algal turf, providing grazing for abundant herbivorous reef fish, particularly Scarids (parrotfishes) and Acanthurids (surgeonfishes). Below this, the reef slope generally dropped gently to depths ranging from 3 to 15m. Coral growth was dominated by the branched hard corals (Acropora and Pocillopora) and massive hard corals particularly Porites. These provided a live hard coral cover that typically ranged between 30 and 65%. However at some sites including Tolka island, live hard cover reached up to 90%. Exceptional cover was also reported for the northern lagoon at Grand Ile by local fishers and dive operators. The base of the slope gives way to a wide platform with a sandy bottom and coral rubble at most locations where the bottom slopes more gently seawards over some distance to a depth of 50m or more. This area is partly covered in coral rubble, scattered coral rock mounds and pinnacles occur providing suitable hard substrate for the growth of some sparsely distributed corals, mostly Favia sp., Acropora spp. and Porites spp. The percent cover of dead hard corals ranged from 1-5% of total substrate at all surveyed locations. Coral mortality did not appear to be recent and the physical damage was minimal. However, at some locations such as the more sheltered Grande Ile Khor Angar, physical damage in terms of anchor damage was observed. Mortality in this area may be due to either fresh water or sediment influx from the mangrove area. Most dead coral was covered with a thin layer of algae. Abundant patterns of coral juvenile recruitment was observed but not quantified during the course work of this project. Soft coral communities were present at all sites with percent cover representing less than 10%. These were dominated by Xenia species. Seagrass beds and mangroves were not recorded at any of the survey stations around the Sept Frères. Details pertaining to specific islands are noted below and related graphs attached in Appendix 11. Grande Ile The shoreline around the island is mainly rocky however, two sandy beaches occur at the northern and southern sides. The island is fringed with well-developed coral communities with a diverse coral assemblage dominated by Porites and Acropora. The percent cover of dead coral colonies (7%) in Grande Ile was the highest value recorded within the survey region. Coral colonies showed signs of physical damage and are covered with a thin layer of algae. Anchor damage could be attributed to approximately 50% of coral mortality. Ile de L'est The shoreline is typically rocky apart from a narrow sandy beach at the southern side. The island is fringed by coral reefs with very narrow and discontinuous reef flats. 35

The reef slope dropped gently from 5-15m forming an inclined plateau that was dominated by sandy bottom with large coral patches. The reef slope then dropped at a 45 0 angle, and live coral cover was dominated by hard corals (Porites and Acropora). Mean hard coral cover at Ile de L'est was 44% of the total bottom cover. Soft coral cover was the highest (9%) within the Sept Frères survey area with Xenia species covering 6% of the surveyed transects. Ile de Sud Coral communities were diverse and abundant on the reef slope which inclined steeply at an angle of 60 towards a sandy bottom with coral patches. The reef flat and back reef were less well developed than other islands. Cover by hard coral was dominated by Porites forming massive colonies and branching Acropora colonies. Hard coral cover ranged between 13% and 64% while the mean percentage cover was 47%. Tolka This small island was characterised by a diverse assemblage coral reef communities and was exposed to strong underwater currents. Tolka was fringed with coral reefs with rare occurrence of a reef flat that was typically covered with a thin layer of algae. Coral cover ranged between 32-66% of the total bottom cover and was dominated by Acropora sp. and Porites sp. Soft coral and Xenia had a mean bottom cover of 12%. However, at the western beach, large Acropora communities formed coral gardens with coral cover of more than 90%. Hamra The coral communities fringing the island were diverse and abundant. Hard coral percentage cover exceeded 60% at some locations. At these locations massive Porites was dominant. A shallow structure extended SW from the island. The area was not surveyed but from the supervised classification it appeared similar to other areas with coral communities on rock Boeing The reef slope inclined at 60' between 5-30m and then dropped sharply. The reef slope was dominated by massive coral (Porites). Xenia was also abundant exceeding 20% of the total cover at some locations. Weathered rock around the island was covered with a thin layer of algae with a width ranging between 0.5-2m. 36

Figure 13: Substrate Transect data, Djibouti 2002 37

Figure 14: A map illustrating the distribution of coral habitats around Sept Frères and Ras Siyyan. Dense coral habitats (in red) indicate where the area is dominated by corals. Sparse coral habitats (in pink) indicate areas where corals were found in association with other more dominant groups such as macroalgae or turf algae. 38

4.2.2 SEAGRASSES Seagrasses were not recorded around the islands of Sept Frères. However, abundant seagrasses formed large beds in Ras Siyyan lagoon (Fig. 11) with four different species recorded, two of which were identified: Halodule and Thallasia species. Halodule and Thallasia species were abundant in the middle and the northern sector of the lagoon. Local fishers reported that one particular seagrass species (observed washed up on the shores of the lagoon and consisting of blades up to 8cm long and a few millimetres wide) was an important food source for dugong (Dugong dugon). However, the location of existence of this species of seagrass was not found nor were any sightings made of dugong during this survey. The marine biotope map (Figure 15) shows the extensive seagrass beds in the lagoon to the north of Ras Siyyan which also extend further north. In addition, seagrass beds were present in the lagoon to the south of Ras Siyyan in front of Khor Angar. Figure 15 illustrates the combined area of the two seagrass biotopes. The total area of seagrass beds within the proposed MPA is 3.2 km 2. Figure 15: The distribution of seagrass biotopes (areas highlighted in purple) around Sept Frères and Ras Siyyan.: 39

4.2.3 MANGROVES Two significant mangrove stands occurred within the survey area, surrounding Ras Siyyan lagoon and at Khor Angar (Figs 16,17). Note that stands in this report refer to areas of mangrove consisting of many trees. The Ras Siyyan stand was crescent shaped surrounding the inner lagoon of Ras Siyyan. It consisted of a monospecific stand of Avicennia marina. This stand ranged from 20 180m width and extended for several kms surrounding the lagoon. The occurrence of Rhizophora mucronata in this area was recalled by fishermen, though no record of the species was identified during this survey. The Khor Angar stand fringed the coastline and the inlet of Khor Angar. The Khor Angar wetland was covered by mangrove vegetation of the two species Avicennia marina and Rhizophora mucronata. R. mucronata formed dense stands fringed by a belt of A. marina along the middle intertidal zone. At both sites, mean tree density covering ranged between 400-1700 trees /ha with mean value 790 trees / ha. Mean height for A. marina was estimated to be 3.9m, while mean height for R.mucronata was 6.3m. Maximum height for A. marina was 8m while R. mucronata attained larger and higher crowns reaching heights that exceeded 12m at some locations. Impacts Impacts to mangroves included felling of R. mucronata trees to use poles for sale for construction purposes and browsing by camels. Piles of cleared timber that exceeded 100 poles were recorded at 3 locations within the southern stand. The mangrove areas surrounding Ras Siyyan bay and at Khor Angar were also crucial nursery habitat for fish, evidence for which was found at the rapid sites, illustrated in Figure 16. Mangroves were also found to provide habitat for feeding, roosting and nesting birds, insects, crustaceans and molluscs (Figure 23 and 24 Birds and Invertebrates at Rapid Sites maps). The extent of mangroves was mapped using the Landsat 7ETM+ satellite data (Figure 9 and 17). The proposed MPA will cover a total of 1.12km 2 of mangrove, including the mangrove area within the bay (0.69 km 2 ), and the area to the north of Khor Angar (0.43 km 2 ). The total area of mangrove illustrated in Figure 18 is 2.61km 2 (includes the mangrove south of Khor Angar with an areas of 1.49 km 2 ). 40

Figure 16: Mangroves at Rapid Sites, Djibouti 2002 41

Figure 17: The distribution of mangroves (areas highlighted in green) around Ras Siyyan Djibouti as identified from the Landsat 7ETM satellite image. 42

4.2.4 FISH, CROWN OF THORNS STARFISH, SEA URCHINS, SEA CUCUMBERS, CLAMS AND GASTROPOD COMMUNITIES (See Survey Form 4A in Appendix 5 for full list of species and family groups). The fish assemblages within the Sept Frères / Ras Siyyan region are highly diverse, and include a combination of reef fish species endemic to the Red Sea and Gulf of Aden region (Pilcher & Abdi, 2000; Kemp, 2001). During underwater visual census, the most abundant species encountered across each family were as follows: Serranidae (groupers): Lethrinidae (emperorfish): Lutjanidae (snappers): Haemulidae (sweetlips): Scaridae (parrotfish): Cephlopholus miniata and Epinephulus tuavina Lethrinus mahsena, L. variegates and L. nebulosus Lutjanus ehrenbergi, L. bohar and Macolor niger Plectorinchus gaterinus and P. gibbosus Scarus sordidus, S. frenatus and S. niger Acanthuridae (surgeonfish): Zebrasoma xanthurum, Acanthurus gahhm, A. sohal The most abundant species on the list were lutjanidae. Chaetodontidae, scaridae, serranidae, haemulidae and pomacanthidae (angelfish) then followed in decreasing order (Fig. 18). Acanthaster planci (Crown of Thorns starfish) scars made up 4.4% of belt counts and lastly sea cucumbers made up 1.7% of counts. The remaining 4.5% labelled Other consisted of: A. planci, emperorfish, clams, Napoleon wrasse, Heterocentrotus & Echinometra urchins, barracuda, moray, Trochus, shark and ray, contributing less than 1% each. Figures 19 and 20 illustrate the relative abundance of the different fish groups throughout the study area. Crown of Thorns scars 4% Sea Cucumber 2% Other 5% Sweetlips 8% Angelfish 5% Snapper 40% Grouper 8% Parrotfish 9% Butterflyfish 19% Figure 18: Abundance of fish and animals from fish transects, Djibouti. Other consists of various fish and other animals (clam and urchins). 43

Figure 19: Relative abundance of different fish groups, Djibouti 2002. 44

Figure 20: Relative fish abundance at Rapid Sites, Djibouti 2002 45

Figure 21: Abundance and distribution of fish groups at six sites, Djibouti 2000 46

When total counts of fish excluding other animals are taken into consideration, two sites at Hamra and Boeing show more than double the abundance of the other sites (Figure.22). A pattern of changing abundance across the surveyed reef sites was apparent for lethrinids and haemulids. Coral reef fish families that are principally most abundant at all sites included scarids, chaetodontids, serranids and lutjanids. It should be stressed here that replication levels are too low to make any statistical analysis significant, however observations indicated that populations of fish around the islands were far greater than along the mainland coastline at Ras Siyyan. The mean values for fish distribution and abundance show that serranids, chaetodontids and scarids dominated the fish assemblage, while the occurrence of large monospecific aggregations of the species Plectorinchus gaterinus and Lutjanus ehrenbergi were recorded at over 60% of the surveyed reefs. These species tend to remain relatively inactive in the water column by day and were sometimes found sheltering close to overhangs, or feeding over sand. As such these aggregations are particularly vulnerable to exploitation to the point of overexploitation. Scarids and chaetodontids in general were most abundant along the shallower reef face, while serranids, lethrinids and lutjanids showed increased abundance at deeper reef slope. In addition, various benthic fauna were recorded with similar species found around the islands. Species common to the islands included holothurians (sea cucumbers), echinoids (sea urchins: Echinometra was found around the islands and Heterocentrotus along the Ras Siyyan coastline) and tridacna (clams) (Figure 22). Figure 22: Distribution and abundance of benthic fauna, Djibouti 2002. 47

4.2.4.1 Sector 1 Ras Siyyan coastline Observations along the Ras Siyyan coastline showed the lowest overall fish abundance of all surveyed reefs (Figure 21). There was no evidence of A. planci nor of fish aggregations. The most abundant fish family amongst those surveyed was butterflyfish and small sized groupers, the latter observed in moderate numbers. High abundance of the other benthic fauna (sea cucumbers and urchins) were also recorded (Figure 22). 4.2.4.2 Sector 2 the islands Grand Ile Reef fish community structure was diverse in this location (high abundance of chaetodontids, scarids and serranids) in addition to pelagic fishes (chanids, carcharinids, carangids and scombrids). Ile de l est The most abundant species recorded belonged to the butterflyfishes. High diversity of these species was recorded (8 species). Lutjanus ehrenbergi occurred in large numbers. Ile de Sud A Large aggregation of Lutjanus ehrenbergi exceeding 150 fish was recorded. Groupers, parrotfishes and butterflyfish were the most abundant of the indicator species included in this survey. Angelfishes (Pomacanthus) were recorded in this island to have the highest abundance within the Sept Frères / Ras Siyyan surveyed locations. Tolka Diverse assemblages of coral reef fish, marine invertebrates, marine mammals and pelagic fishes were observed. The coral reef fish communities showed high abundance compared to other surveyed sites in the Sept Frères / Ras Siyyan proposed MPA. A large aggregation of Lutjanus ehrenbergi was recorded sheltering on the reefs of this island. Also, a large number of juvenile butterflyfish and surgeonfish were observed. Plectorinchus gaterinus occurred in large numbers and were the most abundant fish found around cave entrances and under coral patches. Hamra Large numbers of the species Plectorinchus gaterinus (> 40 fishes) and Lutjanus ehrenbergi (> 200 fishes) were recorded to aggregate on the reefs of this island. The most abundant fish family were butterflyfish. Groupers and parrotfish were less abundant than in other surveyed locations. 4.2.5 CROWN OF THORNS STARFISH A large number of A. planci were recorded around all of the islands of Sept Frères (Figs 22,23). Aggregations of up to 400 animals occurred at some islands and they were observed feeding on hard corals during the day. The mean population size structure of A. planci was 34 cm with a maximum diameter of 55cm. However, several juvenile starfish < 10cm diameter were also recorded. This likely represented an outbreak, even if 48

currently fairly small. A continued growth in the A. planci population would potentially be detrimental to the coral communities in the area. Specific observations and locations: A. planci were recorded in large numbers around the eastern side of Grand Ile with aggregations of 40-100 individuals recorded at four locations around the island A. planci was present in large numbers with between 80-100 recorded at eastern lagoon and the southern headland of Ile de l est. Observations of 10 scars per 100x10m transect indicated a population outbreak. A. planci were observed in large aggregations (50-60 individuals) at two locations around Ile du Sud. They were observed to be feeding on hard corals, typically Favia, Porites and Acropora table coral. At Tolka, A. planci were present in large numbers and feeding during the day (usually they feed at night) and scars to corals from the effect of feeding by A. planci were notably high. An estimated number of 400-600 animals could be present. 49

Figure 23: Distribution and abundance of A. planci (COTs) scars at Rapid Sites, Djibouti 2002 50

Figure 24: Distribution and abundance of A. planci (COTs) scars at Quick Sites, Djibouti 2002 51

4.2.6 TURTLES Two significant turtle nesting sites were found. One at Ile de l Est and one on the northern tip of Ras Siyyan in a secluded beach area (Figure 25 Turtle pits at rapid sites map). Another important turtle nesting site was found on the north-eastern side of Grand Isle (12 28 18.7 N, 043 25 51.4 E) during a reconnaissance survey (Kemp, 2001). Of particular note there was that those nests, along with the extensive high cover and fragile corals in the same bay made that beach of critical conservation management importance. The relatively few nesting sites found on the mainland during this survey are likely due to the high exposure to scavengers, dogs and human activity. Figure 25: Distribution and abundance of turtle pits at Rapid Sites, Djibouti 2002. 52

4.2.7 BIRDS The coast of Ras Siyyan is important for osprey (Pandion haliaetus), and Goliath Heron (Ardea goliath) among various other herons, several species of tern and gull, and sandpipers and crab plovers (Dromas ardeola). Several hundred terns and crab plovers were observed in the western area of the bay. Four pelicans identified as Pelecanus rufescens were also observed daily on the northern section of Ras Siyyan headland. Table 5 summarises the species, number and location of observed birds. Figure 26 summarises bird density and distribution throughout the study area. Table 5. Bird species identified Common and scientific name Osprey (Pandion haliaetus) Brown booby (Sula leucogaster) Hemprich's gull (Larus hemprichi) Pink-backed Pelican (Pelecanus rufescens) White cheeked tern (Sterna repressa) Goliath heron (Ardea goliath) Reef heron (Egretta gularis) Crab plover (Dromas ardeola) Other small beige/brown plovers Range of individuals seen Location seen 1 On islands and Ras Siyyan mainland coast 100 s On islands 100 s On islands and Ras Siyyan mainland coast 1-4 On islands, N side of Ras Siyyan and in mangroves at Khor Angar 100 s On islands 1 On islands and in mangroves at Khor Angar 10 s On islands and Ras Siyyan mainland coast 100 s Along western shore of Ras Siyyan lagoon 10 s Along shore of Ras Siyyan lagoon 53

Figure 26: Bird density and distribution from Rapid Site surveys, Djibouti 2002. 54

4.2.8 CETACEANS Dolphins were sighted daily travelling back and forth along the mainland coastal area or around the islands. Numbers of individuals ranged from small groups of a few animals to larger groups of 14+ animals. Three species were identified during the study period (Table 6). Two dolphin skulls (old) were found on the shores of Ras Siyyan lagoon, one each on the eastern and western shores. Table 6. Species and distribution of dolphin seen in the Ras Siyyan and Iles des Sept Frères area, Djibouti June 2002. Species Minimum and maximum no. of individuals seen at any one time Location seen Spinner dolphin (Stenella longironstris) Bottlenose dolphin (Tursiops truncatus) Humpback dolphin (Sousa chinensis) 2-10 Ras Siyyan coastal waters 14 Ras Siyyan and Iles des Sept Frères coastal waters 2-4 Ras Siyyan coastal waters 4.2.9 DUGONG Although local fishers reported the presence of dugong in the northern area of Ras Siyyan lagoon, no sightings occurred during this study and fishers described the animals as extremely shy of humans and in general rare. 4.2.10 SHARK Only one shark, a lemon shark (Negaprion brevirostris) was observed during transect monitoring at Ile de l est. Anecdotal evidence from fishers and the local dive operator reported a drastic reduction in the shark population over the last decade which had previously been abundant and widespread throughout the islands and off the coast of Ras Siyyan. The cause for this was reportedly due to the illegal shark fin fishery, primarily carried out by Yemenese fishers. Recent patrolling and vigilance around the islands by the French Navy over the last few months was reportedly helping to deter the illegal shark fishery. 4.2.11 TERRESTRIAL INVERTEBRATES, MAMMALS AND REPTILES Large numbers of invertebrates were found particularly in the mangrove areas surrounding Ras Siyyan lagoon and at Khor Angar (Figure 27). 55

Figure 27: Distribution and abundance of invertebrates at Rapid Sites, Djibouti 2002. 56

Small numbers of up to 7 camels were seen at any one time on the mainland. These were most often feeding on the mangroves around Ras Siyyan lagoon and Khor Angar, or moving between the two locations along the beach. Many camel tracks, camel droppings and camel resting spots were observed within the vicinity of the mangroves, particularly around the western side of Ras Siyyan lagoon and at Khor Angar. Gazelle (between 1 to 6 individuals of the same species, but not identified) were also seen along the coastal area between Khor Angar and Ras Siyyan. Many tracks were observed on the western side of Ras Siyyan lagoon inland. Tracks of dog were also found on the western side of Ras Siyyan lagoon and one individual seen. Fishers reported that a species of snake inhabited the island of Horod le Rhale (Ile de l est), the only island thought to have snakes and that these fed on birds eggs. This information was unable to be confirmed during this study. 4.2.12 OTHER OCEANIC MEGAFAUNA Rare and highly charismatic megafauna including whale sharks (Rhincodon typus) and manta rays (Manta birostris) in addition to humphead wrasse (Cheilinus undulates), sharks (Carcharhinidae), turtles and dolphins (Delphinidae) occur in the area surrounding Ras Siyyan and the Iles des Sept Frères. All were seen with the exception of whale sharks and manta rays during the course of this survey. Groups of dolphin (on average 6-8 individuals) were often sighted several times daily throughout the study period. Megafauna are of particular ecological value as well as in terms of attractions to recreational visitors. 4.3 REPORT ON THE INTERTIDAL ENVIRONMENT (See Appendix 16). 4.3.1 EXECUTIVE SUMMARY The objective of the intertidal survey of the proposed regional MPA at iles des Sept Frères and Ras Syan is to assess the status of the habitats located in the intertidal zone for the purpose of developing a site specific Master Plan there. The main habitats are mangroves, and sandy and muddy shores. During this survey the infrastructure has not been developed in the intertidal zone of the MPA. Only few fishermen huts exist there. The species, which live and reproduce in the intertidal zone, are marine turtles, crustaceans, and other species. These key habitats and key species are currently threatened by on going human activities such as camel grazing and hunting which were previously identified by the SAP project document. Marine turtle are hunted by fishermen living inside and outside the MPA and mangroves are exploited unsustainably by herders. Conservation programs for turtles and mangroves have been proposed as well as recommendations for the site-specific master plan. In the mangrove conservation proposal, new ways of exploiting mangrove trees and leaves by herders have been proposed. This conservation plan needs a great deal of awareness so that local communities adopt it. Potential sites, for the development of eco-tourism have also been identified. In addition, revenue-generating activities have been proposed in this report for fishermen and herders engaged in degradation of key habitats and hunting endangered species inside the MPA. 57

4.4 REPORT ON THE SOCIOECONOMIC ENVIRONMENT (See Appendix 16). In addition to the report in appendix 16, the consultants N. Barker and N. Galal gathered the following information: 4.4.1 CENTRES OF POPULATION The sole concentration of people in the study area was that resident in the Khor Angar area at the village of Tourbawad Hori. All six islands were uninhabited although evidence of temporary camps and military activity was present. 4.4.2 SIZE OF POPULATION The population in the study area was estimated to be in the region of 4,000 inhabitants, 200 of which were permanently resident at the village of Tourbawad Hori. Much of the population consisted of nomadic herders or fishers who set up temporary camps throughout the region. No herders were seen during this study, but a population of some 30 fishers consisting of Djiboutians and Yemenese were present in the Ras Siyyan area. During this study, temporary camps were present on the mainland of Ras Siyyan on the eastern and western shores of the lagoon as well as on a rocky outcrop of approximately 80m 2 situated in the middle of the lagoon (N12 28.170, E43 18.477). 4.4.3 OCCUPATIONS Economic and agricultural activities in the study area included artisanal fishing in the lagoon and waters surrounding the islands, camel grazing at mangrove stands around the lagoon and at Khor Angar, and collection of mangrove wood poles for construction and export to Yemen. The latter activity was recorded in the Khor Angar mangrove area. Trading between the local population and the Yemenese was apparent in the Ras Siyyan area and occurred on a daily basis. Other activities in the area included those of the military, both national and international, based on land and at sea. 4.5 REPORT ON THE FISHERIES (See Appendix 16). 4.6 HUMAN IMPACTS 4.6.1 CONSTRUCTION The village of Tourbawad Hori in the Khor Angar area was the only permanent settlement in the study area. Several mall (maximum 4x4m) temporary shelters were present on the mainland along the eastern and western shores of Ras Siyyan lagoon made of mangrove poles. None of the islands had permanent constructions but did have remains of temporary shelters in the form of rocks piled up to provide shelter from the wind. Similar walls of stones, remains of military activity on the islands were also evident. 58

4.6.2 FISHING/COLLECTING Artisanal fishing occurred throughout the Ras Siyyan area but during the survey period was particularly concentrated in the lagoon where fishers were targeting a species of mullet, the Bluespot mullet (Valamugil seheli). Lobster pots were also observed opposite Khor Angar placed on the fringing reef. Fishing material was found at the following sites: 4 permanent ropes tied to the beach at Tolka island used for securing fishing boats. Discarded fishing nets were found at Tolka (x2), Grand Ile (x1) and Ras Siyyan lagoon (x4). Discarded fishing lines were found at Hamra (x4) and on the reef opposite Khor Angar. Discarded fishing rope found at Ras Siyyan. Evidence of probable by-catch was also found at one of the islands and on the mainland at Ras Siyyan. One Spotted Eagle Ray (Aetobatus narinari), less than two days old was found on the beach of Kadda Dabali (Grand Ile) approximately 1m diameter. Three dolphin skulls were found on the mainland, one on the eastern and two on the western shore of Ras Siyyan lagoon. A whole turtle shell of approximately 40cm in diameter and part of another turtle shell was also found on the western side of Ras Siyyan lagoon. 4.6.3 POLLUTION Pollution in the surrounding area of the lagoon primarily in the form of plastic and glass bottles as well as discarded fishing nets. Most of such garbage was found at the tip of Ras Siyyan and other selected spots particularly in the vicinity of temporary shelters and other evidence of human activity (Figure 28). 59

Figure 28: Distribution and abundance of litter at Rapid Sites, Djibouti 2002. 61