UBC Faculty Research and Publications

Fisheries Catch Reconstructions in the Western Indian Ocean, 1950-2010 Le Manach, Frédéric; Pauly, D. (Daniel) 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Le_Manach_F_et_al_Fisheries_catch_reconstructions_western.pdf [ 7.25MB ]
JSON: 52383-1.0354315.json
JSON-LD: 52383-1.0354315-ld.json
RDF/XML (Pretty): 52383-1.0354315-rdf.xml
RDF/JSON: 52383-1.0354315-rdf.json
Turtle: 52383-1.0354315-turtle.txt
N-Triples: 52383-1.0354315-rdf-ntriples.txt
Original Record: 52383-1.0354315-source.json
Full Text

Full Text

ISSN 1198-6727NumberVolumeFisheries Centre Research Reports2015Fisheries CatCh reConstruCtions in the Western indian oCean, 1950-201023 2ISSN 1198-6727 Fisheries Centre, University of British Columbia, CanadaFisheries Centre Research ReportsFisheries Catch Reconstructions in the Western Indian Ocean,  1950-20102015 Volume 23 Number 2Edited byFisheries Centre Research Reports 23(2)ii + 161 pages © published 2015 byThe Fisheries Centre,University of British Columbia2202 Main MallVancouver, B.C., Canada, V6T 1Z4 ISSN 1198-6727 Frédéric Le Manach and Daniel PaulyContentFisheries Centre researCh reports are abstraCted in the Fao aquatiC sCienCes and Fisheries abstraCts (asFa)issn 1198-6727  Fisheries Centre Research Reports 23(2)ii + 161 pages © Fisheries Centre, University of British Columbia, 2015Fisheries Centre Research Reports 23(2)2015Edited byFrédéric Le Manach and Daniel PaulyA Research Report from the Fisheries Centre at UBCForeword iReconstructing Catches for the Union of the Comoros: Uniting Historical Sources of Catch Data for Ngazidja, Ndzuwani and Mwali from 1950–2010 1Beau Doherty, Melissa Hauzer and Frédéric Le Manach Le Développement Soutenu de Pêcheries Artisanales : Reconstruction des Captures Marines à Djibouti de 1950 à 2010 [With an extended English abstract] 13Mathieu Colléter, Ahmed Darar Djibril, Gilles Hosch, Pierre Labrosse, Yann Yvergniaux, Frédéric Le Manach and Daniel PaulyFirst Estimate of Unreported Catch in the French Îles Éparses, 1950-2010 27Frédéric Le Manach and Daniel PaulyTentative Reconstruction of Kenya's Marine Fisheries Catch, 1950–2010 37Frédéric Le Manach, Caroline A. Abunge, Timothy R. McClanahan and Daniel PaulyReconstructing Domestic Marine Fisheries in Mayotte from 1950–2010 53Beau Doherty, Johanna Herfaut, Frédéric Le Manach, Sarah Harper and Dirk ZellerMarine Fisheries in Mozambique: Catches Updated to 2010 and Taxonomic Disaggregation 67Beau Doherty, Margaret M. McBride, Atanásio J. Brito, Frédéric Le Manach, Lizette Sousa, Isabel Chauca and Dirk ZellerReconstruction of the Domestic and Distant-Water Fisheries Catch of La Réunion (France), 1950–2010  83Frédéric Le Manach, Pascal Bach, Léo Barret, David Guyomard, Pierre-Gildas Fleury, Philippe S. Sabarros and Daniel PaulyArtisanal Fisheries in the World's Second Largest Tuna Fishing Ground — Reconstruction of the Seychelles' Marine Fisheries Catch, 1950–2010 99Frédéric Le Manach, Pascal Bach, Léa Boistol, Jan Robinson and Daniel PaulyFailed State: Reconstruction of Domestic Fisheries Catches in Somalia 1950–2010 111Lo Persson, Alasdair Lindop, Sarah Harper, Kyrstn Zylich and Dirk ZellerSouth Africa’s Marine Fisheries Catches (1950–2010) 129Sebastian Baust, Lydia Teh, Sarah Harper and Dirk ZellerAn Update of the Reconstructed Marine Fisheries Catches of Tanzania with Taxonomic Breakdown 151Elise Bultel, Beau Doherty, Adam Herman, Frédéric Le Manach and Dirk ZelleriThis Fisheries Centre Research Report presents catch reconstructions for the fisheries of eleven entities of the western Indian Ocean, ranging from mighty South Africa in the South to the tiny Djibouti in the North, and from the uninhabited Îles Eparses — the scattered islands in and near the Mozambique Channel — to the densely populated Comoros Islands, and from the flourishing democracy of the Seychelles to the failed state that is Somalia. Only Madagascar and Mauritius are not included here — but their catches were reconstructed earlier.1 The contrasts that these wide ranges of climate, population, and governance styles represent are not necessarily related as one would expect with the state of marine fisheries resources of these various entities. Thus, Somalia and Djibouti may have, overall, the least impacted coastal fish stocks of the region, while stocks along the Indian Ocean Coast of South Africa, a country which boasts a considerable research and management capacity, are generally in dire straits.Although the 'national' catch reconstructions presented here emphasize the coastal and thus domestic catches of Western Indian Ocean countries, the attention of their fisheries ministries is often directed elsewhere, i.e., at the tuna fisheries that have made the Western Indian Ocean a bonanza for European (mainly Spanish and French) and East Asian fishing fleets.2These fisheries are very lucrative, and the Western Indian Ocean countries that permit tuna vessels to operate in their Exclusive Economic Zones (EEZs) get a share — albeit small — from access fees. But they should not forget their domestic fisheries because they are the ones that ensure — or at least contribute — to the (sea)food security of their populations. The eleven chapters in this report document that this contribution to seafood security is much higher than previously assumed, but also that it is sharply declining in several countries.Thus, this report suggests that it is time to devote more attention to coastal fisheries and to rehabilitate them, e.g., as achieved in Southern Kenya, where strong, positive interactions between governments, scientists and fishers enabled the banning of gears not suitable for sustainable reef fisheries, and the establishment of marine reserves. Hopefully, initiatives such as these can be duplicated throughout the region.The EditorsParis and VancouverFebruary 20151 For Madagascar: Le Manach F, Gough C, Humber F, Harper S and Zeller D (2011) Reconstruction of total marine fisheries catches for Madagascar (1950–2008). Pp. 21–37 In Harper S and Zeller D (eds.), Fisheries catch reconstructions: islands, part II. Fisheries Centre Research Reports 19(4). University of British Columbia, Vancouver (Canada).For Mauritius: Boistol L, Harper S, Booth S and Zeller D (2011) Reconstruction of marine fisheries catches for Mauritius and its outer islands, 1950–2008. Pp. 39–61 In Harper S and Zeller D (eds.), Fisheries catch reconstructions: Islands, Part II. Fisheries Centre Research Reports 19(4). University of British Columbia, Vancouver (Canada).2 Le Manach F, Chavance P, Cisneros-Montemayor AM, Lindop A, Padilla A, Schiller L, Zeller D and Pauly D (in press) Global catches of large pelagic fishes, with emphasis on the high seas. In Pauly D and Zeller D (eds.), Global atlas of marine fisheries: ecosystem impacts and analysis. Island Press, Washington, DC (USA).ForeWordFisheries catch reconstruction for the Comoros — Doherty et al. 1reConstruCting CatChes For the union oF the Comoros: uniting historiCal sourCes oF CatCh data For ngazidja, ndzuWani and mWali From 1950–2010*Beau Doherty,1 Melissa Hauzer2 and Frédéric Le Manach1,3† 1 Sea Around Us, Fisheries Centre, University of British Columbia, 2202 Main Mall, Vancouver V6T 1Z4, Canada2 Department of Geography, University of Victoria, Victoria V8W 3P5, Canada3 Institut de Recherche pour le Développement, UMR212 Ecosystèmes Marins Exploités, Avenue Jean Monnet, CS 30171, 34203 Sète cedex, France† Current address: BLOOM Association, 77 rue du Faubourg Saint-Denis, 75010 Paris, Franceb.doherty@fisheries.ubc.ca; mdhauzer@gmail.com; fredericlemanach@bloomassociation.org abstraCtComorian fisheries consist of a small-scale boat fleet of pirogues and motor boats operated by men as well as shore-based fishing by women, both of which have few catch statistics. We compiled historical data on catch rates and the number of boats in the fleet from both grey literature and national statistics, and used these to reconstruct a time series of boat-based catch from 1950–2010. We also estimate catch by women fishers by extrapolating per-capita catch rates from a recently published study on the island of Ngazidja. Catches increased slowly from 1,000 t in 1950 to around 5,000 t in 1980, after which catches increased rapidly due to the increasing number of motorized vessels and the use of FADs offshore. The size of the fleet has grown rapidly since the 1990s and catch estimates are highest from 2005–2010 at around 19,000 t∙year-1. Overall, the reconstructed catches are 1.3 times the figures reported to and by the FAO for the Indian Ocean. Total reconstructed catches consist primarily of Thunnus albacares (yellowfin tuna), Katsuwonus pelamis (skipjack tuna), Sardinella spp. (sardinellas) and Engraulidae (anchovies).introduCtionThe Union of the Comoros (referred throughout as 'the Comoros') is an archipelago in the northern Mozambique channel of the Western Indian Ocean that is composed of three islands: Ngazidja (or Grande Comore), Ndzuwani (or Anjouan) and Mwali (or Mohéli; Figure 1). Fisheries in the Comoros consist of a small-scale pirogues and fibreglass motor boat (locally known as barques or vedettes) fleet as well as shore-based subsistence fishing by women. Until the 1980s the boat fleet was almost exclusively non-motorized pirogues using mostly handlines (de San 1983). Catches by this fleet increased in the 1980s, due to the importation of fibreglass motorboats and the use of anchored Fish Aggregating Devices (a-FADs) for fishing further offshore (Cayré 1991; Anon. 2013). The high season fishing months have historically occurred between November and February when tunas migrate around the islands; catches in other months are generally lower (Van Nierop 1985; James 1988).The fishing conditions on the three islands vary. Ngazidja is surrounded by a narrow coral reef extending  about 500 m from shore (Fourmanoir 1954). The reef is generally not very deep, nor is it followed by a large shelf with productive fishing areas for reef species. Fishing here has historically targeted large pelagic species (e.g., sharks,1 tuna, billfish and dolphinfish) in areas about 5–15 km offshore (Fourmanoir 1954). The fishing conditions off  Ndzuwani are comparable to Ngazidja, except the productivity of the coral reefs is higher (Fourmanoir 1954). Fishermen from Ndzuwani and Ngazidja frequently fish the waters near Mwali, perhaps even more heavily than local residents (James 1988). Ndzuwani fishers also regularly fish in Mayotte's waters (Maggiorani et al. 1993; Doherty et al. this volume).* Cite as: Doherty B, Hauzer M and Le Manach F (2015) Reconstructing catches for the Union of the Comoros: uniting historical sources of catch data for Ngazidja, Ndzuwani and Mwali from 1950–2010. Pp. 1–11 In Le Manach F and Pauly D (eds.) Fisheries catch reconstructions in the Western Indian Ocean, 1950–2010. Fisheries Centre Research Reports 23(2). Fisheries Centre, University of British Columbia [ISSN 1198–6727]. 1 Historically there was a targeted fishery for sharks (Fourmanoir 1954), but fishing effort now targets more tuna and it does not appear that there is a fishery targeting sharks specifically. When caught, sharks are finned for the Chinese market and the meat is consumed locally (Kiszka et al. 2008).EEZ BoundaryShelf 0 100 km±NgazidjaMwaliNdzuwaniFigure 1.  Map of the Comoros showing the islands of Ngazidja (Grande Comore), Ndzuwani (Anjouan) and Mwali (Mohéli), the extent of the Exclusive Economic Zone (EEZ), as well as the continental shelf (darker blue). 2Mwali is the least densely populated island, and its continental shelf extends about 10 km from the coast and harbours much more productive reef fisheries (Fourmanoir 1954). Thus, fisheries in Mwali have historically targeted more reef species (e.g., snappers, groupers and emperors). A marine park was established around the southern half of the island in 2001, extending from the high tide mark to 100 m depth and covering over 400 km2 (Granek and Brown 2005).Fishing in the Comoros is mostly day trips, as no vessels have refrigerating units (de San 1983; James 1988) and crews are small, usually 1–2 people per boat (WIOMSA 2011). The average crew and boat sizes are larger in Ngazidja, as these vessels often fish further offshore (WIOMSA 2011). Catches are usually landed on shore and sold on the same day (Anon. 2013). There are no restrictions on the amount of fish that may be landed, but fishers rarely catch more than they can consume or sell in one day (Hauzer et al. 2013a).A survey conducted by the Direction Nationale des Ressources Halieutiques (National Fisheries Department) in 1994 provided the only comprehensive assessment of catch by the small-scale boat fleet, and Hauzer et al. (2013a) provided the only catch estimates for women fishers. Fishers reported declines in catch abundances, mean size and changes in species composition over the last two decades (Hauzer et al. 2013b), but there was a lack of official catch statistics to verify these trends. Other studies during the 1950–2010 period provided estimates of catch rates and the number of boats in the fleets, and were used to reconstruct a time-series of boat-based catch. The size of the fleet has grown rapidly since the 1990s, but data collection and catch statistics has remained limited.The purpose of this report was to estimate the total marine fisheries catches for the Comoros from 1950 to 2010, using FAO data as the baseline. Reconstructed estimates were compared with FAO landings in an attempt to identify unreported sources of catch and increase transparency in Comorian catch statistics.the Comoros' Fisheries and reConstruCtion methodsBoat-based catchWe compiled catch and effort data from grey literature and unpublished datasets for select years with available data from 1950–2010. These data were treated as anchor points and linear interpolation of catch rates, and the number of fishing boats were used to estimate gaps in the data between anchor points.Boat time-seriesTable 1 summarizes the boat counts that were available for the small-scale fishing fleets since 1950. In the early 1950s, Fourmanoir (1954) estimated that there were 130 pirogues in the southern villages of Ngazidja. Moal (1962) counted 213 boats in these same villages, representing a ratio of 0.61:1 for boats in 1954 compared to those in 1962. We assumed that the same increase occurred proportionally on other areas of Ngazidja, Ndzuwani and Mwali and applied this ratio to boat counts by Moal (1962) to estimate boat numbers in 1954. There was a large discrepancy in the number of boats recorded in the 1993 and 1994 national survey data. Since we found no explanation to justify this, we used the average number of boats from 1993–1994.The boat anchor points in Table 1 were converted to a boat per-capita2 rate for each boat type and linear interpolation was used to estimate boats per-capita for years without boat data. This provided a boat per capita time-series from 2 Population statistics for 1960–2012 were available from the World Bank database (http://databank.worldbank.org, Accessed: 06/06/2013) and for 1950 and 1955 from the United Nations database (http://data.un.org, Accessed: 06/06/2013). Missing years in the 1950s were linearly interpolated.Table 1.  Anchor points for the number of boats in the Comoros for various years between 1950–2011.Year Pirogues Motor boatsb SourcesNon-motorized MotorizedaNgazidja Ndzuwani Mwali Ngazidja Ndzuwani Mwali Ngazidja Ndzuwani Mwali1954 566 123 36 - - - - - - Fourmanoir (1954)c1962 928 201d 59 - - - - - - Moal (1962)1979 1,455 970 194 45 30 6 - - - Faharoudine (1979)1983 - - - - - - 15e 25 10e Van Nierop (1985)1987 1,500 1,200 300 - - - 11 18 6 James (1988)1993 2,012 1,391 242 107 120 79 250 92 69 Unpub. data, 1994 survey, Direction National des Ressources Halieutiques1994 1,748 1,505 247 87 80 54 109 77 392011 1,888 1,864 227 23 25 209 802 708 18 Unpub. data, 2011 boat census, Direction National des Ressources Halieutiquesfa Includes Fedawa I.b Includes barques, vedettes, Fedawa II, Yamaha G18, and Japawa). Classification of boat categories are based on boat size, capacity, and horse power (Lablache-Carrara and Laloë 1993; Aboulhalik 1998).c Multiplied 1962 boat numbers by 0.61.d Missing boat data for 3 communities (Vouani, Pomoni and Moya) was estimated using the median value from other communities in Ndzuwani.e Estimated the distribution of vedettes in 1983 for Ngazidja and Mwali. There were originally 50 vedettes supplied (James 1988), 25 of which were originally on Ndzuwani (Van Nierop 1985). f Obtained through SWIOFP database ( catch reconstruction for the Comoros — Doherty et al. 31950–2010 that was used to estimate a boat time-series by multiplying per capita boat rates by annual populations on each island (Figure 2).Catch rates time seriesWhere available, estimates of annual catch were divided by total boats on the island to estimate an annual catch rate. Otherwise, daily catch rates were multiplied by the number of trips per year to estimate the annual catch for different boat types (Table 2). Early observations of small-scale fisheries in the Comoros archipelago were well-documented by Fourmanoir (1954). These included catch rates of 50 kg·trip-1 for pirogues fishing twice per week on the southern offshore banks of Mwali, and an average annual catch rate of 5.2 t·pirogue-1.  Pirogues fishing every day on the interior reef had an average catch rate of 10 kg·trip-1 and an annual catch rate of 2.6 t·pirogue-1, assuming trips occurred 5 times per week (Table 2).During this time, a normal week of pelagic fishing by 70 pirogues off Ngazidja landed: 1,500 kg of sharks (Isurus glaucus and Carcharinus longimanus), ten yellowfin tuna (Thunnus albacares) with an average weight of 20 kg, two dolphinfish (Coryphaena hippurus) with an average weight of nine kilos, and four Indo-Pacific sailfish (Istiophorus platypterus). Average lengths of landed sailfish on Ngazidja were 2.9 m (Fourmanoir 1954), and we estimate their average weight as 50 kg per fish using length-weight relationships from FishBase (www.fishbase.org). This yielded a total of 1,918 kg of pelagic fish by 70 boats in one week, or an average weekly catch of 27.4 kg per pirogue. Based on the overall catch composition in the 1994 survey data (unpub. data, Direction Nationale des Ressources Halieutiques), we assumed an additional 20% of annual catches were composed of yellowtail barracuda (Sphyraena flavicauda), oilfish (Ruvettues pretiosus), small pelagics and other reef fish (e.g., Lethrinidae and Carangidae), also documented in catches by Fourmanoir (1954). Linear interpolation between catch rates was used to fill in gaps. We used the same rates for all three islands where island specific rates were not available. We had no catch rates beyond 1994; however, the majority of boat fishers interviewed by Hauzer et al. (2013a) reported declines in catch abundance and mean fish sizes over the last 20 years. This was not surprising given the large increase in motorized vessels in the small-scale fishing fleet during this time (Figure 2). We assumed catch rates declined by 50% between 1994 and 2010 (Figure 3).1950 1960 1970 1980 1990 2000 2010 Numer of boats  0      500   1 000   1 500   2 000  NgazidjaMwaliNdzuwaniNgazidjaMwaliNdzuwani1979 1984 1989 1994 1999 2004 2009  0      306090120NgazidjaMwaliNdzuwani1983 1988 1993 1998 2003 2008  0      600700800900  200300400500100YearABCFigure 2.  Estimated annual number of A) non-motorized pirogues from 1950–2010, B) motorized pirogues from 1979–2010, and C) motor boats from 1983–2010. Solid circles represent anchor points described in Table 1. 4Shore-based fishing by womenHauzer et al. (2013a) provided the only estimates of catch by women fishers in the Comoros. Based on data collected from interviews in 2009–2010, the annual catch by women was estimated for three communities on  Ngazidja (Table 3). Each community had 80 women fishers and catch estimates for each village ranged from 40–99 tonnes due to the differences in catch rates and the number of full-time fishers. There are another seven communities on the island consisting of women fishers, one of which contains only ten fishers (Hauzer et al. 2013a). We estimated that the other six villages also contained approximately ten women fishers, based on the second author's personal experience. Extrapolated to the entire island of Ngazidja, this yielded an estimate of 255 tonnes for 2010. We converted these 255 tonnes to a shore-based catch per-capita and assumed the same per-capita catch rate for the islands of Mwali and Ndzuwani in 2010.  All women fishers interviewed by Hauzer et al. (2013a) reported declines in catch abundance in the last ten years. We therefore assumed a higher shore-based catch per-capita in 2000, increasing the 2010 per-capita rate by 25%. We used linear interpolation to derive per-capita catch rates from 2001–2009 and maintained the 2000 rate from 1950–1999.Taxonomic and sectoral breakdownWe maintained the same annual taxonomic compositions for the boat-based reconstructed catch as what was reported to the FAO for the 1970–2010 period. Data reported to the FAO from 1950–1969 had poor taxonomic detail, with 20–70% of the catch reported as 'marine fishes nei' (i.e., unidentified marine fish). To improve the taxonomic breakdown prior to 1970, we reallocated catches of unidentified fish to other taxonomic groups based on the catch composition in the early 1970s, using the average 1971–1973 breakdown from FAO. The assumption here was that new taxa reported in 1971–1973 FAO landings (e.g., Carangidae, Engraulidae, Istiophoridae and Scombriae) were reflective of improved taxonomic reporting rather than new fish species targeted by fisheries. This retained 7% Table 3.  2010 annual catch estimates for full-time and part-time women fishers on Ngazidja (derived from Hauzer et al. 2013a). Community Annual catch (t) Average catch per year Number of fishers % full timeFull time Part time Full time Part timeChindini 40.2 0.9 0.4 10 70 0.13Hantsindzi 59.7 0.9 0.5 50 30 0.63Mitsamiouli 98.8 1.4 0.7 60 20 0.75Seven other villages 8.1  1.1a  0.5a 5 5  0.50aa Estimated as average value from the three communities above.Table 2.  Summary of annual catch rate estimates from 1954 to 1994.Year Boat type Daily catch rate (kg·boat-1) Trips per year Annual catch rate (t·boat-1)aSources and commentsNgazidja1954 Non-motorized pirogue 34.3a - 1.8 Fourmanoir (1954)1962 - - 1.5 Moal (1962)1978 - - 1.4 de San (1983)1994 - - 1.7 Unpub. data, 1994 survey, Direction Nationale des Ressource HalieutiquesMotorized pirogue - - 12.2Motor boat - 23.7Ndzuwani1962 Non-motorized pirogue - - 3.5 Moal (1962)1979 - - 1.4 de San (1983) 1983–1984 39.7 (high season), 21.8b (low season)104.0 3.0 Van Nierop (1985), James (1988). Annual catch rate assumes 60% of trips occur in low season and 40% of trips occur in high season (Van Nierop 1985; James 1988)Motorized pirogue 69.0 (high season), 38.0b (low season)104.0 5.2Motor boat 252.4 (high season), 138.9b (low season)122.0 22.61986–1987 Non-motorized pirogue 20.0 104.0 2.1 James (1988)Motor boat 250.0 122.0 30.51994 Non-motorized pirogue - - 1.4 Unpub. data, 1994 survey, Direction Nationale des Ressources HalieutiquesMotorized pirogue - - 8.0Motor boat - - 15.0Mwali1954 Non-motorized pirogue 50 (offshore) and 10 (inshore)104 (offshore) and 260 (inshore)c3.9 Fourmanoir (1954). Average of inshore and offshore rate1962 - - 5.1 Moal (1962)1979 - - 2.1 Faharoudine (1979), de San (1983)1994 - - 0.8 Unpub. data, 1994 survey, Direction Nationale des Ressources HalieutiquesMotorized pirogue - - 4.7Motor boat - - 9.8a Numbers reflect weekly catch rates observed in October; number of trips per week unknown. b Low-season rate estimated as 55% of high season rate based on ratios from James (1988). c  Fourmanoir (1954) indicates these fishers fish every day; here we assume an average of 5 trips per week throughout the year (Van Nierop 1985).Fisheries catch reconstruction for the Comoros — Doherty et al. 5of annual catches as 'unidentified marine fish' and reassigned the remainder as anchovies, bigeye tuna, Carangidae, Indian mackerel, Indo-Pacific sailfish, kawakawa, skipjack tuna, swordfish and yellowfin tuna.The Sea Around Us uses the following fishing sectors in its global catch database: 'industrial' (i.e., large-scale commercial), 'artisanal' (i.e., small-scale commercial), 'subsistence' (i.e., small-scale non-commercial with primary purpose being self- or family-consumption), and 'recreational' (i.e., small-scale non-commercial with the primary purpose being pleasure). For the purposes of the Sea Around Us database, the small-scale shore fishing and boat-based catches were divided into artisanal and subsistence components. Species of higher value, such as tunas (Thunnus spp. and Euthynnus spp.), billfishes (Istiophoridae) and lobsters, were assumed to be primarily sold commercially and thus 80% of this catch was allocated to the 'artisanal' sector. Fish species sold at lower prices, such as small pelagics (Clupeiformes), and marine molluscs, were assumed to be primarily used for take-home consumption and 80% of these catches were allocated as 'subsistence'. For other species where the distinction was not obvious, such as jacks (Carangidae), mackerels (Auxis spp., Rastrelliger spp., Scomberomorus spp.), and other unidentified marine fish, we used an even split, allocating 50% to each small-scale sector. All shore-based catch was allocated as subsistence.Flags of ConvenienceFAO landing data (FishStat 2014) also included catches from the Central Eastern Atlantic (FAO Area 34) from 2007–2012. These landings were composed mostly of pelagic species (primarily Clupeidae, Engraulidae, and Carangidae) as well as a small amount (2%) of demersals. However, as there were no records of any distant water Comorian fishing fleets, these catches were not considered domestic.Rather, we suspected that these catches were from industrial fishing vessels fishing the high seas using the Comoros as a flag of convenience. As a matter of fact, the Comoros has been previously identified as a potential flag of convenience state for high seas fishing (Gianni and Simpson 2005; Anon. 2013). The FAO fishing vessels finder database (www.fao.org/figis/vrmf/finder; Accessed: June 26, 2014) listed six foreign vessels (Table 4) that have been registered with the Comoros flag between 2004 and 2012. Given the absence of any other information we assumed that these countries were responsible for the FAO reported landings in the Atlantic. We allocated 89% of the pelagic catch to Lithuania 1950 1960 1970 1980 1990 2000 2010 Catch per pirogue or boat (t) 0    3456NgazidjaMwaliNdzuwani1979 1984 1989 1994 1999 2004 2009  0    81012141983 1988 1993 1998 2003 2008  0      2025303551015YearBC21A642MwaliMwaliNgazidjaNgazidjaNdzuwaniNdzuwaniFigure 3.  Estimated annual catch rates (t·boat-1) for A) non-motorized pirogues from 1950–2010, B) motorized pirogues from 1979–2010, and C) motor boats from 1983–2010. Solid circles represent anchor points described in Table 2. 6and 11% to Netherlands based on the proportion of total tonnage by vessels from these countries using midwater otter trawls. Demersal catches were reallocated to France, the only country with vessels using bottom trawls.Although we could not confirm that France, Lithuania and Netherlands were responsible for these catches, we are confident that these catches were not from the Comoros fishing fleets and our reallocation is, therefore, more informative than what is currently in the FAO database.Foreign fisheriesIndustrial longliners from Japan have fished in Comorian waters since at least the late 1950s (Moal 1962). Although there were no formal agreements at this time, Japan has contributed funds and equipment (e.g., fiberglass vessels, fishing nets) to develop the Comoros' fisheries over the years (de San 1983; James 1988; Lablache-Carrara and Laloë 1993). Formal agreements with the European Union have been signed in recent years, which allowed up to 45 tuna seiners and 25 longliners from France, Spain, Italy and Portugal to fish for tuna in the Comoros EEZ (Anon. 2013; Eckstein 2014).resultsOverall, the total reconstructed catches from 1950–2010 were nearly 516,000 t, 96% of which were from the small-scale boat fleet, and 4% of which were from shore-based fishing by women (Figure 4A). Catches increased slowly from 1,000 t in 1950 to around 5,000 t in 1980, after which catch volumes increased rapidly due to the increasing number of motorized vessels and the use of offshore a-FADs. The size of the fleet has grown rapidly since the 1990s and despite decreasing catch rates, catch estimates were the highest from 2005–2010 at around 19,000 t∙year-1. Overall, the reconstructed catches were 1.3 times the landings reported to FAO in the Indian Ocean. Total reconstructed catches consisted primarily of yellowfin tuna, skipjack tuna, sardinellas, and anchovies (Figure 4B). The sectoral assignments suggested that 'artisanal' and 'subsistence' catches accounted for 61% and 39% of total reconstructed catches from 1950–2010, respectively (Figure 4A).disCussionThe overall discrepancy between the reconstructed domestic catches and the data reported to FAO was mainly due to an increase in catch since 1995, which contributed 54% of the total reconstructed catch (and were 95,000 t higher than what was reported to FAO). In estimating catches for these years, we assumed a 50% decline in catch rates since the 1994 survey by the Direction Nationale des Ressources Halieutiques,which is the only comprehensive assessment of small-scale fisheries in the Comoros. Table 4.  Fishing vessels registered with the Comoros flag in the FAO fishing vessel finder database.Country Boat name Period Gear Length (m) Gross tonnage Lithuania ARAS-1 2009, 2012 Midwater otter trawl 104 4,378IRVINGA 2011–2013 105 4,407BALANDIS 2011, 2012 109 5,953KOVAS 118 5,979The Netherlands OCEAAN VII 2012 90 2,624France LA MADONE 2 2004 Bottom trawl and dredging 11 160 Boat-based, artisanal5 10 15 20 Boat-based, subsistenceShore-based, subsistenceAB0 1950 1960 1970 1980 1990 2000 2010 Catch (thousand tonnes) Year 5 10 15 20 Yellowfin tunaSkipjack tunaSardinesAnchoviesOthersJacksBillfishesOther tunas and mackerelsFigure 4.  Reconstructed catches by A) sector and B) major taxa. 'Others' includes lobsters, molluscs, sharks, rays and other unidentified marine fish. See details in Appendix Table A1 and Appendix Table A2.Fisheries catch reconstruction for the Comoros — Doherty et al. 7The 2012 dataset produced by FAO (2012; i.e., the dataset used here) showed that catches have steadily decreased from 1994 to around 11,000 t∙year-1 from 2001–2010. This was in stark contrast to the 2010 dataset (FAO 2012), where catches increased from 1994 to 20,500 t∙year-1 in 2008–2010. As we found no catch statistics for this period, the reasons for this change in the FAO data remains unknown. What we do know is that the number of motor boats in the Comoros increased from around 300 in 1994 to about 1,700 in 2011 (Figure 2). Catches in the 2012 FAO data declined by 23% over this same period and, if accurate, would correspond to about a 70% decline in annual catch per boat since 1994 (Figure 5).Interviews with fishermen confirmed that there has been a decrease in mean fish sizes and perceived catch abundance over the last 20 years (Hauzer et al. 2013b), but by how much we do not know. In these interviews 62% of motorized boat fishers and 55% of pirogue fishers reported declines in fish sizes. Similarly, 50% of motorized boat fishers and 75% of pirogue fishers reported declines in catch abundance. Given that this perception was not consistent among fishers, Hauzer et al. (2013b) suspected that the declines in catch were not drastic. Furthermore, catch figures remained reasonably high in recent years, averaging 22 kg∙day-1 for pirogues and 110 kg∙day-1 for motor boats (Hauzer et al. 2013b). Thus, we think that our assumption of a 50% decline may well be conservative, and catches over this period may actually be higher. Not surprisingly, sensitivity analysis of this assumption shows that catches in the last 15 years would vary considerably depending on the decline in catch rates assumed; thus this is a major source of uncertainty in our estimates (Figure 5). We were not able to provide estimates of uncertainty for the reconstructed totals, as error estimates are unavailable for catch statistics used, including those reported by the FAO.For most of the 1950–1994 period, we found the 2012 FAO dataset to be a reasonable estimate of boat-based catches. The reconstructed estimates here provide an alternative, but show a similar trend to FAO data. The reconstructed catches yielded per-capita consumption rates of 6–15 kg∙person-1∙year-1 from 1950–1979. These are low for an island country with few other protein sources. However, it is known that throughout the 1950s-1970s, fishing did not satisfy local consumption requirements and large amounts of salted fish were imported from Madagascar and Zanzibar (Fourmanoir 1954; Moal 1962; Faharoudine 1979; Meyer et al. 2006). Further studies are necessary to improve confidence in our results, notably with regards to the shore-based activities conducted by women fishers. In this study we extrapolated estimates by Hauzer et al. (2013b) for the island of Ngazidja to estimate catches for all of the Comoros and using population data estimated historical catches by this sector. This information provides a preliminary estimate of the scale of these catches, which could be improved through specific studies on the islands of Mwali and Ndzuwani. The importance of such activities for food security and livelihoods is increasingly recognized (Harper et al. 2013; Anon. 2014; Kleiber et al. 2014). Thus, further research is required to better understand the species most affected by these fisheries and well as their social and economical impact. We found few data on the species composition of catches in the Comoros and much of the data reported to FAO from 1950–1969 was recorded as unidentified marine fish. We attempted to improve the taxonomic detail of these catches, by disaggregating them based on more detailed information in FAO data in the early 1970s. Information on major species caught by fisheries in 1950s (Fourmanoir 1954) and the 1994 national statistics could be used in future efforts to improve Comorian catch statistics and may provide valuable information of changes in species composition over time. For example, approximately 80% of catches on Ngazidja observed by Fourmanoir (1954) over a one-week period were composed of sharks, whereas they accounted for less than 1% of annual catches on Ngazidja in 1994 and fishermen reported that sharks are now rarely seen in catches (M. Hauzer, unpub. data). It is clear that tunas are now the main target species for offshore pelagic fisheries, but sharks likely accounted for a much larger proportion of catches in earlier years and this is not reflected in our estimates. Groupers (Plectropomus pessuliferus,3 Epinephelus merra, Variola louti), snappers (Lutjanus argentimaculatus, Aprion virescens) and emperors (Lethrinus nebulosus, L. olivaceus4) were commonly caught in Mwali reef fisheries in the 1950s (Fourmanoir 1954), but we found no species specific catch statistics for reef fisheries in recent years to compare these with. Interviews with fishermen suggested that some species that were once common are now rarely seen, 3 Plectropomus maculatus listed in Fourmanoir (1954) is a likely misidentification (Froese and Pauly 2014).4 Lethrinus miniatus listed in Fourmanoir (1954) is a likely misidentification (Froese and Pauly 2014).0 5 10 15 20 25 1950 1960 1970 1980 1990 2000 2010 Catch (thousand tonnes) Year Reported to FAOFigure 5.  Reconstructed boat-based catch compared to other sources of catch data. Solid circles represent estimates observed in the grey literature (Moal 1962; de San 1983; Van Nierop 1985; James 1988; Amoriggi 2010) and the 1994 survey by the Direction National des Ressources Halieutiques. Dashed (75% decline) and dotted (25% decline) lines show sensitivity analysis for different assumptions about the decline in catch rates between 1994–2010.  8while other species have completely disappeared from catches (M. Hauzer, unpub. data). The lack of detailed catch statistics for the Comoros makes it difficult to assess the magnitude of such changes, their causes and their impacts on fisheries and marine ecosystems It is critical for fisheries management that the Comoros dedicates more resources to accurately recording fisheries statistics (Pauly et al. 2013). Other than the 1994 national statistics and a few recently published studies (e.g., Hauzer et al. 2013a,b), data that exist are mostly from grey literature and often based on brief observations of the fishery in select regions. It is unlikely that we will ever know the 'true' historic catches of small-scale fisheries in the Comoros, but we hope this work may serve as a starting point to account for unreported catch statistics, improve transparency in fisheries data, and provide a resource of historical information for Comorian fisheries. If there are additional data that were unavailable to us, we hope they may be used to improve this work and we welcome contributions from other researchers to improve this database, which will be made available via the Sea Around Us website.aknoWledgementsThe authors of this report would like to acknowledge the support of the Sea Around Us, a collaboration between the University of British Columbia and The Pew Charitable Trusts. BD and FLM also thank the Paul G. Allen Family Foundation for support.reFerenCesAboulhalik FM (1998) Marine science country profiles — Comores. Intergovernmental Oceanographic Commission, Western Indian Ocean marine science association (WIOMSA). 35 p.Amoriggi S (2010) Union des Comores — Appui au développement de la transformation des produits agricoles aux Comores. Fonds international de développement agricole (FIDA), Rome (Italy).Anon. (2013) Contrat Spécifique n°4 — Évaluation rétrospective et prospective du protocole de l'accord de partenariat dans le secteur de la pêche entre l'Union européenne et l'Union des Comores. Contrat Cadre MARE/2011/01 — Évaluation et analyse d'impacts — Lot 3 — Évaluations rétrospectives et prospectives relatives à la dimension internationale de la Politique Commune des Pêches, Cofrépêche, Marine Resources Assessment Group (MRAG), Poseidon Aquatic Resource Management Ltd, and Nordenfjeldske Development Services (NFDS), Brussels (Belgium). 111 p.Anon. (2014) Sustainable fisheries and aquaculture for food security and nutrition. A report by the High Level Panel of Experts on Food Security and Nutrition, Rome (Italy). 118 p.Cayré P (1991) Behaviour of yellowfin tuna (Thunnus albacares) and skipjack tuna (Katsuwonous pelamis) around fish aggregating devices (FADs) in the Comoros Islands as determined by ultrasonic tagging. Aquatic Living Resources 4: 1–12.de San M (1983) Profil de la pêche artisanale aux Comores. South West Indian Ocean Fishery Project SWIOP/RAF/79/065/WP/6/82, FAO, Moroni. 22 p.Doherty B, Herfaut J, Le Manach F, Harper S and Zeller D (this volume) Reconstructing domestic marine fisheries in Mayotte from 1950–2010. Eckstein A (2014) MEPs green-light three fisheries agreements. Europolitics, edition of April 23, 2014.Faharoudine AH (1979) Country statement on the fisheries situation in the Comoros. pp. 45–47 In Report of the FAO/IOP workshop on the fishery resources of the Western Indian Ocean south of the Equator. Indian Ocean Fishery Commission, Indian Ocean Programme, Development Report No. 45. FAO, Mahé (Seychelles).FAO (2012) FishstatJ Capture Production 1950–2010. Food and Agriculture Organization of the United Nations (FAO), Rome (Italy).FAO (2014) FishstatJ Capture Production 1950–2012. Food and Agriculture Organization of the United Nations (FAO), Rome (Italy).Fourmanoir P (1954) Ichthyologie et pêche aux Comores. Mémoires de l'Institut Scientifique de Madagascar Série A — Tome IX. 187–239 p.Froese R and Pauly D, editors (2014) Fishbase (www.fishbase.org).Gianni M and Simpson W (2005) The changing nature of high seas fishing: how flags of convenience provide cover for illegal, unreported and unregulated fishing. Australian Department of Agriculture, Fisheries and Forestry, International Transport Workers' Federation, and WWF International. 83 p.Granek EF and Brown MA (2005) Co-management to Marine Conservation in Mohéli, Comoros Islands. Conservation Biology: 1724–1732.Harper S, Zeller D, Hauzer M, Pauly D and Sumaila UR (2013) Women and fisheries: contribution to food security and local economies. Marine Policy 39: 56–63.Hauzer M, Dearden P and Murray G (2013a) The fisherwomen of Ngazidja island, Comoros: fisheries livelihoods, impacts, and implications for management. Fisheries Research 140: 28–35. Hauzer M, Dearden P and Murray G (2013b) The effectiveness of community-based governance of small-scale fisheries, Ngazidja island, Comoros. Marine Policy 38: 346–354.Fisheries catch reconstruction for the Comoros — Doherty et al. 9James JRW (1988) Summary of the fisheries and resources information for the Comores. In Sanders MJ, Sparre P and Venema SC (eds.), Proceedings of the workshop on the assessment of the fishery resources in the southwest Indian Ocean. RAF/79/065/WP/41/88/E. FAO/UNDP, Albion (Mauritius).Kiszka J, Wickel J, Jamon A and Layssac K (2008) Note sur l'exploitation et la capture accidentelle des requins autour de la Grande Comore (Union des Comores), Canal de Mozambique. Document interne — Groupe de Recherche sur les Requins — MAYSHARK. 2 p. Kleiber D, Harris LM, Vincent AC and Rochet M-J (2014) Improving fisheries estimates by including women's catch in the Central Philippines. Canadian Journal of Fisheries and Aquatic Sciences 71(5): 656–664.Lablache-Carrara G and Laloë F (1993) Plan d'exécution d'un système d'échantillonnage des captures aux Comores. 46 p.Maggiorani JM, Minet JP and Taquet M (1993) Situation de la pêche artisanale à Mayotte en 1992–93 — Résultats de l'enquête halieutique (mars 1992 — février 1993) et comparison avec l'enquête halieutique de 1988–89. IFREMER, Collectivité Territoriale de Mayotte — Direction de l'Agriculture et de la Forêt, Service des Pêches. 34 p.Meyer J, Nelson D, Black D and TANGO International (2006) Comoros: comprehensive food security and vulnerability analysis (CFSVA). World Food Programme (WFP), Vulnerability Analysis and Mapping Branch (ODAV). 62 p.Moal RA (1962) La pêche aux Comores. Ministère d'état chargé départements et territoires d'outre-mer, France. 78 p.Pauly D (2013) Fisheries: does catch reflect abundance? Nature 494(7437): 303–306.Van Nierop M (1985) Rapport préliminaire sur la performance des embarcations de pêche sur l'île d'Anjouan aux Comores. RAF/79/065/WP/20/85, SWIOP. Available at: http://www.fao.org/docrep/field/279543.htm#F007 [Accessed: July 16, 2012].WIOMSA (2011) Migrant fishers and fishing in the Western Indian Ocean: socio-economic dynamics and implications for management. Final report of commissioned research project MASMA/CR/2008/02. 158 p. 10Appendix Table A1.  Total reconstructed catch compared to official data reported to FAO.Year Reconstructed Reported to FAO Shore-based  Boat-based Total Artisanal  Subsistence 1950  142  653  347 1,142 8351951  145  659  431 1,235 9351952  148  634  549 1,331 9181953  152  739  542 1,433 8181954  155  796  584 1,535 8181955  158  975  518 1,651 8351956  161  1,052  559 1,772 8351957  164  1,091  643 1,898 7171958  167  1,166  688 2,021 7171959  170  1,143  839 2,152 8171960  173  1,221  896 2,290 8171961  176  1,171  1,087 2,434 8171962  179  1,285  1,115 2,579 9171963  182  1,344  1,165 2,691 9171964  185  1,403  1,217 2,805 9171965  189  1,318  1,407 2,914 9171966  193  1,212  1,617 3,022 8751967  197  1,258  1,678 3,133 8751968  202  1,425  1,619 3,246 1,2351969  206  1,470  1,671 3,347 1,2351970  211  1,420  1,815 3,446 1,6621971  215  1,647  1,672 3,534 2,4701972  220  1,832  1,564 3,616 2,8791973  225  1,985  1,486 3,696 3,2871974  230  2,074  1,472 3,776 4,0471975  237  2,115  1,517 3,869 4,7561976  245  2,315  1,412 3,972 4,8641977  255  2,357  1,470 4,082 5,6211978  265  2,378  1,408 4,051 6,0271979  275  2,821  1,610 4,706 6,4861980  285  3,131  1,732 5,148 6,9521981  295  3,428  1,863 5,586 7,4601982  304  3,711  1,990 6,005 7,9751983  312  6,684  3,540 10,536 8,4941984  321  6,774  3,543 10,638 9,0001985  330  6,136  3,175 9,641 9,5161986  339  5,444  2,773 8,556 9,9711987  348  5,543  2,757 8,648 10,3791988  357  6,349  3,151 9,857 10,9141989  366  7,345  3,282 10,993 10,7521990  375  7,804  3,855 12,034 11,2521991  385  8,321  4,291 12,997 11,5521992  394  8,853  4,597 13,844 12,5911993  403  9,624  4,807 14,834 12,7581994  413  9,191  4,510 14,114 13,5371995  423  9,484  4,756 14,663 13,1091996  434  9,838  4,934 15,206 12,6961997  445  10,110  5,170 15,725 12,5761998  456  10,368  5,358 16,182 12,3171999  468  10,701  5,475 16,644 11,8182000  480  10,770  5,826 17,076 12,0032001  482  11,077  5,895 17,454 11,4252002  485  11,246  6,069 17,800 11,1782003  487  11,314  6,289 18,090 11,0532004  489  11,263  6,587 18,339 10,9872005  491  11,383  6,661 18,535 10,7382006  492  11,393  6,784 18,669 10,4642007  494  11,102  7,156 18,752 10,7242008  495  10,921  7,352 18,768 11,0932009  496  10,803  7,414 18,713 10,8252010  496  10,621  7,449 18,566 10,540Fisheries catch reconstruction for the Comoros — Doherty et al. 11Appendix Table A2.  Total reconstructed catch by taxa.Year Yellowfin tuna Skipjack tuna Sardines Anchovies Jacks Others1950  477  99  120  44  69  334 1951  463  98  233  43  67  332 1952  396  95  386  47  74  333 1953  479  117  313  58  90  376 1954  516  126  337  62  96  397 1955  701  159  179  66  102  445 1956  754  173  193  71  110  470 1957  735  186  242  89  138  508 1958  783  201  258  95  148  535 1959  735  189  485  89  139  515 1960  782  204  518  95  148  542 1961  714  183  829  81  126  500 1962  788  209  785  96  150  552 1963  821  221  821  101  157  572 1964  857  231  857  105  163  592 1965  759  206  1,188  87  136  538 1966  673  175  1,616  48  74  436 1967  695  185  1,677  49  77  450 1968  822  225  1,479  73  113  535 1969  845  235  1,526  75  116  550 1970  732  202  1,557  195  195  565 1971  865  249  1,209  134  269  808 1972  1,072  319  1,062  118  236  809 1973  1,235  379  950  106  211  815 1974  1,249  397  876  88  219  947 1975  1,269  432  916  76  191  984 1976  1,423  552  766  77  192  963 1977  1,386  607  817  68  170  1,033 1978  1,383  675  754  63  157  1,019 1979  1,607  870  840  68  178  1,142 1980  1,742  1,039  881  70  189  1,227 1981  1,855  1,213  936  71  199  1,312 1982  1,950  1,392  987  79  207  1,391 1983  3,404  2,643  1,733  144  361  2,250 1984  3,345  2,816  1,720  138  355  2,265 1985  2,932  2,668  1,526  117  313  2,084 1986  2,525  2,479  1,327  99  264  1,862 1987  2,498  2,638  1,288  96  256  1,873 1988  2,760  3,134  1,471  113  287  2,092 1989  3,283  3,756  988  494  494  1,978 1990  3,441  3,938  1,036  1,036  518  2,065 1991  3,626  4,149  1,092  1,310  546  2,274 1992  4,000  3,906  1,068  1,068  534  3,268 1993  4,712  3,972  1,131  984  566  3,470 1994  4,643  3,408  1,012  881  506  3,664 1995  4,799  3,615  1,086  978  543  3,642 1996  4,941  3,828  1,163  1,047  570  3,656 1997  4,954  3,951  1,215  1,093  571  3,941 1998  4,987  4,103  1,277  1,149  600  4,066 1999  5,112  4,347  1,300  1,163  616  4,105 2000  4,928  4,339  1,452  1,313  691  4,353 2001  5,040  4,605  1,486  1,263  743  4,318 2002  4,993  4,743  1,549  1,317  775  4,424 2003  4,861  4,816  1,593  1,433  796  4,591 2004  4,682  4,850  1,868  1,543  812  4,583 2005  4,554  4,954  1,680  1,596 1,008  4,742 2006  4,410  5,053  1,737  1,650 1,042  4,776 2007  4,033  4,890  1,958  1,873 1,107  4,891 2008  3,621  4,667  1,894  1,812 1,071  5,704 2009  3,413  4,705  1,935  1,851 1,094  5,714 2010  3,185  4,728  1,972  1,886 1,114  5,680 Fisheries catch reconstruction for Djibouti — Colléter et al. 13le développement soutenu de pêCheries artisanales : reConstruCtion des Captures marines à djibouti de 1950 à 2010*Mathieu Colléter,1,2 Ahmed Darar Djibril,3 Gilles Hosch,4 Pierre Labrosse,5 Yann Yvergniaux,6 Frédéric Le Manach1,7† and Daniel Pauly11 Sea Around Us, Fisheries Centre, University of British Columbia, 2202 Main Mall, Vancouver V6T 1Z4, Canada2 Agrocampus Ouest, UMR985 ESE Ecologie et santé des écosystèmes, Rennes, France3 Direction de la Pêche, Ministère de l'Agriculture, de l'Elevage et de la Mer, Chargé des Ressources Hydrauliques, Djibouti-ville, Djibouti4 Fisheries Planning & Management, PO Box 862, L-2018 Luxembourg5 Mission pour la recherche et la technologie, Haut-commissariat de la République en Nouvelle-Calédonie, Nouméa, France6 SmartFish Programme, Indian Ocean Commission, Ebène, Mauritius7 Institut de Recherche pour le Développement, UMR212 Ecosystèmes Marins Exploités, Avenue Jean Monnet, CS 30171, 34203 Sète cedex, France† Current address: BLOOM Association, 77 rue du Faubourg Saint-Denis, 75010 Paris, Francem.colleter@fisheries.ubc.ca; djidarar@hotmail.com; hosch@pt.lu; labrosse.pierre@gmail.com; yann.yvergniaux@coi-ioc.org; fredericlemanach@bloomassociation.org; d.pauly@fisheries.ubc.caextended abstraCtThis study is part of the Sea Around Us and consists of a 'reconstruction' of the likely total fisheries catch (i.e., domestic and foreign catch) made in Djibouti's waters from 1950 to 2010, as well as catches by fishers from Djibouti in foreign waters. This reconstructed time-series contrasts with official catches reported to the Food and Agriculture Organization of the United Nations (FAO) by Djibouti, which were found to be incomplete and misleading. Indeed, failed to account for several sectors such as discards, subsistence and recreational fisheries, or illegal catches by foreign fleets. The reconstruction of Djibouti's marine fisheries catches over the 1950–2010 period overlapped with the development of Djibouti's artisanal sector and allowed us to account for the aforementioned missing sectors. A thorough bibliographic research on fisheries in Djibouti was carried out, and 'anchor points' required for estimating historical catches were then identified (Pauly 1998), similar to other reconstructions made around the world (see, e.g., Zeller and Pauly 2007; Zeller and Harper 2009; Harper and Zeller 2012; Harper et al. 2013). We separated our analysis into three sections: Djibouti's catch within its own Exclusive Economic Zone (EEZ), Djibouti's catch outside its EEZ, and finally, foreign catches within Djibouti's EEZ.Our results show that FAO data contained several inconsistencies. Artisanal catches for the pre-independence period appeared to be too high given the low number of fishers and fishing practices at that time. For the post-independence period, our reconstruction was based on data provided by the Département de la Pêche (Department of Fisheries) — deemed to better reflect Djibouti's fisheries trends — which also substantially differed from those published by FAO. Our reconstruction also included crude estimates for previously unaccounted sectors, which enabled us to produce a more realistic picture of the overall catch within Djibouti's EEZ and in foreign waters. Most notably, we identified that 25% of the total artisanal catches were not declared. In addition, we identified and estimated a small subsistence fishery that was never accounted for in official statistics, similarly to other sectors such as an holothurian fishery, artisanal discards, and a recreational fishery by tourists and residents. We also estimated catches made outside Djibouti's EEZ, mostly in Somalia. Illegal foreign fisheries, mostly from Yemen (but also Somalia to a lower extent), were also estimated. Finally, we also improved the taxonomic composition of catches during the early time-period using FAO's breakdown in more recent years. Overall, our total reconstructed catches (mostly composed of inshore species such as Serranidae, Lutjanidae, Carangidae, and Lethrinidae) are roughly similar to total catches reported to FAO, but annual catches are wildly different, especially in the earlier time-period. Before the independence in 1977, we estimated that catches increased from 130 t∙year-1 to almost 300 t∙year-1, due to an increase in the number of fishers. Afterward, the annual catch quickly increased to 1,000 t by the end of the 1980s, thanks to the development of a real fisheries vision by the national authorities. Catches steeply decreased in the mid-1990s due to the Civil War, but then steadily increased again to reach around 1,300 t∙year-1 in the 2000s. This report also provide some evidence of overexploitation, and we recommend to improve the data collection scheme and statistical framework in Djibouti, in order to better monitor domestic and foreign fisheries and thus ensure a sustainable use of marine resources.*  Cite as: Colléter M, Darar Djibril, A, Hosch G, Labrosse P, Yvergniaux Y, Le Manach F and Pauly D (2015) Le développement soutenu de pêcheries artisanales : reconstruction des captures marines à Djibouti de 1950 à 2010. Pp. 13–25 In Le Manach F and Pauly D (eds.) Fisheries catch reconstructions in the Western Indian Ocean, 1950–2010. Fisheries Centre Research Reports 23(2). Fisheries Centre, University of British Columbia [ISSN 1198–6727]. 14résuméCette étude menée dans le cadre du Sea Around Us avait pour objectif de reconstruire les captures effectuées dans les eaux de Djibouti et pas les pêcheurs djiboutiens en dehors des eaux nationales, afin de les comparer aux captures déclarées à la FAO. Cette étude a recoupé le développement de la pêcherie artisanale djiboutienne, et nos résultats montrent que les captures artisanales déclarées au programme des Nations Unies pour l'alimentation et l'agriculture (FAO) avant l'indépendance sont trop élevées par rapport au nombre de pêcheurs et aux pratiques de l'époque. Pour la période post-indépendance, nous avons estimé que l'équivalent de 25% des captures artisanales déclarées ne l'étaient pas. Nous avons également mis en évidence une pêche de subsistance non-déclarée, mais relativement faible (2% du total déclaré). D'autres secteurs ont aussi été inclus, tels que la pêche non-déclarée d'holothuries, les rejets de la pêche artisanale, et la pêche récréative. Les captures totales effectuées dans la ZEE djiboutienne sont en fin de compte relativement proches de celles déclarées par Djibouti à la FAO, mais les captures annuelles sont très différentes. Nous avons également estimé les captures artisanales djiboutiennes faites en dehors de la ZEE nationale (principalement en Somalie). Enfin, nous avons estimé les pêches illégales étrangères en provenance du Yémen, mais aussi de Somalie. Ces résultats montrent qu'il est aujourd'hui nécessaire de mieux surveiller les activités de pêche à Djibouti afin d'améliorer la collecte des données et ainsi garantir une gestion durable des ressources.introduCtionLa République de Djibouti ('Djibouti' sera utilisé par la suite) est un petit pays d'Afrique de l'Est situé au nord de la Somalie, face au Yémen. Il occupe une position stratégique à la jonction de deux grands ensembles maritimes, le golfe d'Aden et la Mer Rouge, reliés par le détroit de Bab-el-Mandeb (Figure 1). Cette position stratégique entre la Mer Rouge et l'Océan Indien en a fait un point de contrôle essentiel pour les intérêts commerciaux et militaires français avant l'indépendance en 1977 (Devinat 1957). Ces intérêts n'ont pas limités à la France, et on y trouve encore aujourd'hui des bases militaires américaines, japonaises, et autres. L'espace maritime sous juridiction djiboutienne s'étend environ 7 200 km² dont 2 500 km² de plateau continental (parmi les plus modestes du continent africain, avec une largeur moyenne estimée à 8 km ; El Gharbi 1987 ; Bouhlel 1988 ; Hosch 2010). Ce plateau est fortement développé au Nord et au Sud de la bande côtière, et est plus étroit à l'intérieur, atteignant tout au plus 1,5 km de largeur à certains endroits. Il recouvre en grande partie le golfe de Tadjourah, où la grande vallée du rift s'enfonce dans la mer, et la rive sud du golfe s'étire jusque dans les eaux somaliennes, alors que la rive nord se prolonge par une longue plaine (Bouhlel 1988). Des récifs coralliens son présents sur presque toute la bande côtière et le pourtour des îles attenantes, jusqu'à une profondeur de 20–30 m. Au-delà, on observe un fond en pente douce couvert de sable et/ou de vase (Bouhlel 1988 ; Hosch 2010).La structure hydrologique est conditionnée par l'opposition de deux masses d'eau : celle du Golfe d'Aden dans la couche supérieure, et celle plus profonde originaire de la Mer Rouge. La température en surface est d'environ 31°C en été et 26°C en hiver ; en profondeur, la température la plus basse observée était de 15°C vers 600 m (Allain 1974). Le régime des vents influence fortement la situation en surface, et conditionne la profondeur de la thermocline (20–30 m en été et autour de 100 m en hiver), la quantité d'oxygène dissous, ainsi que la salinité. Ces paramètres influent sur le comportement des espèces et les saisons de pêche (Abbes 1985). La saison sèche qui s'étend de juillet à août est caractérisée par des vents (Khamsiin) allant jusqu'à sept sur l'échelle de Beaufort, entrecoupée de tempêtes sporadiques empêchant les sorties en mer. Entre les mois de novembre et mars apparaît souvent une houle allant jusqu'à cinq Beaufort (particulièrement dans le Nord), freinant également les activités de pêche (Künzel et al. 1996a). Les principales saisons de pêche sont donc comprises entre avril-début juin et août-octobre (El Gharbi 1987 ; Künzel et al. 1996a).Les récifs coralliens sont les écosystèmes côtiers les plus représentatifs à Djibouti, mais il existe également un réseau important de lagunes, mangroves et herbiers sous-marins (El Gharbi 1987). Les fonds sont productifs jusqu'à une profondeur de 70 m, la faune se raréfiant au-delà (Allain 1974). L'étendue du plateau continental aux extrémités nord (notamment autour des îles des Sept Frères ; El Gharbi 1987) et sud du pays permet la présence de stocks importants de poissons démersaux. La région nord jouît également de conditions très favorables avec la présence des eaux froides de la mer d'Arabie engendrant une forte production primaire et une abondance des petits pélagiques. De même, dans le golfe de Tadjourah, la remontée de la thermocline pendant l'été est suivie d'une explosion de la production primaire stimulant ainsi la production des stocks pélagiques. La descente de la thermocline en hiver réduit l'activité planctonique et entraîne le départ vers d'autres régions de la majorité des stocks associés (Bouhlel 0 50 km±ShelfEEZ boundaryObockTadjourahDjibouti-villeLimites de l  ZEEPlateau océaniqueÎles dessept frèresÎles MushaFigure 1.  Carte de la ZEE de Djibouti, montrant l'étendue du plateau océanique (bleu foncé) et de la Zone Exclusive Économique (ZEE), ainsi que les principales villes côtières de Djibouti-ville (capitale), Tadjourah et Obock.Fisheries catch reconstruction for Djibouti — Colléter et al. 151988). Ainsi, une grand diversité d'espèces est pêchée (Bouhlel 1988). Les taxons les plus importants sont ceux des grands démersaux nobles (Acanthuridae, Balistidae, Epinephelinae, Haemulidae, Labridae, Lethrinidae, Lutjanidae, Mullidae, Scaridae, et Sparidae) et les grands pélagiques (Carangidae, Carcharhinidae, Coryphaenidae, Istiophoridae, Scombridae, Sphyraenidae, Sphyrnidae, et Xiphiidae). Des stocks de Sepia spp. (seiches), Loligo spp. (calmars), Holothuria spp. et Actinopyga palauensis (holothuries), et crustacés  (crevettes, cigales, langoustes et crabes), présentent également un intérêt (Hosch 2010). Les petits pélagiques (Carangidae, Clupeidae, et Scombridae) et les petits démersaux restent quant à eux majoritairement inexploités par la flotille djiboutienne (Künzel et al. 1996a ; Morgan 2006). Traditionnellement, la pêche a toujours été faible à Djibouti, et le poisson semble peu consommé par la population (Bjoerklund et Walter-Dehnert 1983). La tradition des Afars et des Issas, les ethnies premières de Djibouti, est plutôt pastoraliste que maritime (comme en Somalie ; cf. Persson et al. ce volume). Le développement de cette activité a principalement été dû aux Yéménites qui fréquentaient saisonnièrement la côte africaine et qui, pour certains, se sont installés sur le territoire au moment de sa colonisation par la France. C'est à partir de ces populations yéménites fixées sur le territoire depuis deux ou trois générations que s'est constituée la première communauté de pêcheurs de Djibouti (Moal et Grateau 1967 ; Clouet 1970 ; Rouaud 1997). Par la suite, la pêche s'est développée dans la partie sud du pays majoritairement peuplée par les Issas, rattachés au groupe ethnique Somali et profondément influencés par les Arabes. Les Afars, peuplant majoritairement le Nord du pays, sont restés plus longtemps tournés vers la terre et négociaient des droits de pêche avec les Yéménites venus sur la côte (Clouet 1970). Durant la période coloniale, l'activité de pêche est restée peu développée avec un faible nombre de pêcheurs, peu de moyens, et une volonté du gouvernement français portée en premier lieu sur l'exploration du domaine maritime (Moal et Grateau 1967 ; Allain 1974). Suite au gain de l'indépendance en 1977, le gouvernement djiboutien a souhaité développer cette activité afin d'exploiter pleinement son potentiel halieutique, créer des richesses et améliorer la sécurité alimentaire du pays. Cette volonté s'est traduite par la réalisation de projets de soutien au développement des activités de pêche, financés par plusieurs instances internationales. L'objectif était le développement d'une pêcherie exclusivement artisanale et nationale (Waldstein et Lampe 1988 ; Darar 1994 ; Morgan 2006). La pêche industrielle a donc toujours été formellement interdite au sein des eaux djiboutiennes, les seuls bateaux autorisés à la pêche (inférieurs à 16 m et sans chalut de fond) devant posséder une licence accessible aux seuls citoyens djiboutiens (Künzel et al. 1996a ; Darar et Hosch 2010). De manière générale, la pêche est donc longtemps restée un travail à temps partiel dû aux revenus faibles et moyens de production limités, et a gardé sa nature artisanale, étant pratiquée en zone côtière avec de petites embarcations. Entre 1982 et 1986, plus de 60% de la production nationale étaient réalisés par l'Association Coopérative de Pêche Maritime (ACPM) située à Djibouti-ville. En 1986, 14 patrons pêcheurs sur 165 débarquant à l'ACPM produisaient 56% des captures avec 35% des sorties en mer. Les zones de pêche étaient principalement concentrées au Sud (seulement 2% des captures de l'ACPM dans le Nord), et l'activité à terre était concentrée à Djibouti-ville avec du matériel au potentiel limité (El Gharbi 1987). L'activité s'est par la suite développée et professionnalisée au cours des années 90 et 2000 avec des projets de soutien et l'arrivée de nouveaux opérateurs (Künzel et al. 1996a ; Emerton 1998 ; Hosch 2010). Les moyens de production et équipements se également sont améliorés, permettant d'exploiter les zones au Nord du pays à fort potentiel (El Gharbi 1987), et représentant 77% des captures totales en 2010 (Direction de la Pêche).Les sorties en mer ne dépassent jamais 72 heures, et se font le plus souvent à la journée selon les marées et les vents. Elle a peu changé au cours des années bien qu'ayant connu des progrès technologiques depuis les années 50 (apparition du monofilament et des moteurs hors-bords, par exemple). Les principaux types sont la pêche à la ligne (palangrotte) visant les démersaux, la ligne de traîne visant les pélagiques, et accessoirement le filet maillant (El Gharbi 1987 ; Künzel et al. 1996a). Ils se pratiquent depuis des embarcations, les houris, pouvant contenir deux à trois hommes pour les petits bateaux (six à huit mètres) et cinq pour les plus grands (10–14 mètres). La pêche se pratique également à pied sur le plateau madréporique avec l'utilisation d'éperviers (pour la pêche à la crevette notamment), de filets, et également la pêche en apnée (en particulier pour la pêche à la langouste ; Clouet 1970 ; Moal et Grateau 1967 ; Künzel et al. 1996a). Cette pêche à pied est supposée de faible envergure,et ne  constituerait qu'une petite activité annexe ou de subsistance (Künzel et al. 1996a ; Morgan 2006). Les séries statistiques publiées par l'Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO) sont souvent incomplètes de par leur non prise en compte de plusieurs secteurs tels que la pêche de subsistance et la pêche récréative. Ici, nous proposons de reconstruire les captures déclarées à la FAO par Djibouti depuis 1950 afin d'en améliorer la qualité et la lecture, notamment par la réestimation des secteurs manquants, ainsi que la ré-allocation des captures aux différents secteurs.. matériels et méthodesLes données servant de base au travail présenté ici ont été extraites de la base de données FAO FishstatJ (FAO 2012). Elles correspondent aux données de captures déclarées par Djibouti de 1950 à 2010. Une recherche bibliographique portant sur les pêches dans les eaux djiboutiennes a ensuite été effectuée afin d'en comprendre et compléter le contenu. Des 'points d'ancrage', requis pour l'estimation des captures historiques depuis 1950 (Zeller et Pauly 2007), ont ensuite été identifiés. Nous avons séparé notre analyse en trois composantes : l'étude des captures de Djibouti à l'intérieur de sa ZEE, celle des captures de Djibouti à l'extérieur de sa ZEE, et enfin celle des captures étrangères dans la ZEE djiboutienne. 16Reconstruction des captures djiboutiennes à l'intérieur de la ZEE nationaleCaptures déclarées de la pêche artisanale djiboutienneNous avons dissocié deux périodes pour la reconstruction des captures déclarées de la pêche artisanale djiboutienne : les périodes (i) pré-indépendance de 1950 à 1977, et (ii) post-indépendance de 1978 à 2010. La période pré-indépendance est caractérisée par une très faible disponibilité d'informations. Cependant, l'ensemble des éléments trouvés dépeignent une pêche peu développée avec peu de pêcheurs (Moal et Grateau 1967 ; Moal 1969 ; Clouet 1970), contredisant clairement les données FAO présentant des captures relativement hautes entre 1950 et 1963.1 La période post-indépendance est caractérisée par plus de données, notamment de la Direction de la Pêche (branche du gouvernement chargée du suivi et contrôle des pêches dans le pays).2  • Période pré-indépendanceNous avons reconstruit les captures de cette période en utilisant le nombre de pêcheurs présents dans la ville de Djibouti, à savoir, 107 en 1951 et 150 en 1967 (Moal et Grateau 1967). Ces deux valeurs ont constitué nos points d'ancrage pour le calcul des captures de la pêche artisanale entre 1950 et 1977. Nous disposions également des captures annuelles par pêcheur, calculées entre 1982 et 1986 pour l'ACPM (El Gharbi 1987). Afin d'uniformiser nos méthodes de calcul avec celles développées par El Gharbi (1987) et utilisées pour le calcul des captures de la pêche artisanale entre 1982 et 1986 (reprises par la Direction de la Pêche ; cf point suivant), nous avons considéré que ces deux valeurs du nombre de pêcheurs correspondaient au nombre total de pêcheurs à Djibouti-ville, et divisé ces valeurs par 1,5 (i.e., le nombre moyen de pêcheurs par équipage). De cette manière, nous avons obtenu le nombre de 'patrons pêcheurs', que nous avons par la suite multiplié par une valeur constante de prise annuelle par pêcheur égale à environ 1 285 kg·année-1·pêcheur-1 (i.e, les deux-tiers de la moyenne des captures annuelles par pêcheur sur la période 1982–85).3 Ainsi, nous avons considéré que la prise par pêcheur dans les années 50 et 60 était inférieure de un tiers à celle effectuée au début des années 80, reflétant les progrès techniques substantiels entre les deux périodes et une dynamique plus active de la filière (cf. Introduction).  D'autre part, nous avons également ajouté aux captures obtenues 5% (poids des viscères) et 7,9% (captures dans les autres localités ; i.e., la moitié de la proportion observée en 1986 pour traduire le faible développement de la pêche dans le Nord ; El Gharbi 1987). Nous avons ensuite interpolé linéairement (extrapolé pour l'année 1950) les captures entre nos deux point d'ancrages en 1951 et 1967, puis entre 1967 et la moyenne des captures en 1978–79. Selon plusieurs auteurs, ces captures comprenaient plusieurs espèces telles que les mérous, carangues, barracudas, thons, bonites, maquereaux royaux et langoustes (que l'on retrouve dans les débarquements récents), mais également les mulets, aiguillettes, et sardinelles, très peu ciblés de nos jours (Moal et Grateau 1967 ; Clouet 1970). Il semblerait donc que les espèces débarquées aient changé au cours du temps, les captures se concentrant de plus en plus sur les poissons de grande taille grâce aux progrès technologiques et au développement de la pêche professionnelle (Darar 1994). De plus, plusieurs petits types de pêche existaient à l'époque, tels que la pêche à la nacre (Pinna spp. ; pratiquée principalement par les érythréens sur le récif des îles Musha et dont le marché très limité disparut à la fin des années 70), ou la pêche au corail (vendu aux touristes ; Clouet 1970). Cependant, les informations concernant la composition taxonomique et son évolution au cours du temps restant extrêmement limitées, nous avons décidé de conserver la composition taxonomique présente dans les données rapportées à la FAO de 1983 à 1987, comme pour le début de la période post-indépendance.• Période post-indépendancePour cette période, la Direction de la Pêche nous a fourni la série de données des captures nationales de la pêche artisanale, que nous avons substituée aux données de la FAO (considérées comme moins représentatives/précises) pour le tonnage. La série de la Direction de la Pêche ne comportait cependant aucune résolution taxonomique. De la même manière que pour la période pré-indépendance, nous avons donc utilisé la composition des données déclarées à la FAO de 1983 à 1987.4Secteurs non-déclarésPlusieurs secteurs n'ont jamais été pris en compte dans l'estimation des captures djiboutiennes. Tout d'abord, une partie des captures réalisées par la pêche artisanale djiboutienne était et est toujours non-déclarée à la Direction de la Pêche. Cette pêche artisanale produit également des rejets non comptabilisés. Une pêche de subsistance a également existé depuis longtemps, pratiquée par quelques pêcheurs de la capitale et dans d'autres régions (Morgan 2006 ; Direction de la Pêche, données non publiées). Enfin, la pêche récréative, de plus en plus importante, prélève également nombre d'espèces marines sans suivi adéquat.1 A noter que l'intégralité des données FAO pour la période pré-indépendance est constituée du groupe générique 'marine fishes nei'.2 Les données déclarées par Djibouti ont été corrigées par la FAO pour la période 1983–91 (estimations basées sur 70% de la production officiellement déclarée ; FAO 1991). Il en est de même pour la période 1992–2004, où les données de la FAO correspondent à des estimations dont les méthodes de calcul ou sources ne sont pas précisées (FAO 2012).3 Nous avons exclu l'année 1986 de ces calculs car elle a connu une augmentation de la production à l'ACPM due à une nette amélioration de l'efficacité de pêche (El Gharbi 1987).4 Entre 1979 et 1982, les données publiées par la FAO incluent déjà Panulirus spp.. Ces captures rapportées ont donc été soustraites aux captures ré-estimées à partir de la composition taxonomique de 1983 à 1987.Fisheries catch reconstruction for Djibouti — Colléter et al. 17• La pêche artisanale non-déclaréeCe secteur équivaut à un pourcentage non négligeable des captures déclarées (Künzel et al. 1996a ; Emerton 1998 ; Hosch 2010). Ce phénomène est ancien, et a été causé par un développement lent de la pêche artisanale et de la filière associée (e.g., infrastructures de conservation, commercialisation), et donc de son contrôle (Clouet 1970 ; Abbes 1985 ; Darar 1994 ; Morgan 2006). Les captures non-déclarées sont en partie débarquées et vendues à Djibouti en dehors des circuits contrôlés (e.g., restaurateurs). Elles sont également données aux personnes qui aident lors du débarquement, gardées pour auto-consommation, ou jetées à cause d'une mauvaise conservation par les pêcheurs (dans le cas de l'ACPM, cette dernière cause a été estimée à 5% des captures en 1986, El Gharbi 1987 ; aussi estimée hypothétiquement à 25% des captures artisanales, Emerton 1998). El Gharbi (1987) a inclus une partie de ces pratiques (pour l'ACPM seulement, soit 3% des captures artisanales déclarées) dans ses estimations des captures artisanales nationales reprises par la Direction de la Pêche. De plus, des débarquements de pêcheurs basés à Obock et sur la côte Nord se faisaient et font toujours au Yémen (marché plus attractif, accès au carburant fortement détaxé, et proximité des points de débarquements). Enfin, une grande partie de la pêche se faisait hors contrôle, étant donné que nombre de pêcheurs artisanaux ne possédaient pas de licence. En effet, la mise en place de licences n'a eu lieu qu'en 2008 (Hosch 2010), et sont aujourd'hui délivrées aux propriétaires de bateaux (les pêcheurs non propriétaires se voyant également attribuer des cartes. Les captures de cette pêche artisanale non-déclarée correspondraient à 20–30% des déclarations officielles (e.g., Künzel et al. 1996a). Nous avons donc utilisé une valeur de 25% de données non-déclarées pour l'ensemble de la période 1950–2010.Les espèces concernées par cette pêche non-déclarée semblent similaires à celles dont la capture est déclarée, même s'il semblerait qu'il existe quelques différences. Par exemple, plusieurs groupes seraient plus particulièrement visés : les requins (principalement des Carcharhinidae ; Anon. 2011) et les Mugilidae (pêche à l'épervier au Sud; Direction de la Pêche). Les petits pélagiques (e.g., Clupeidae, principalement Sardinella longiceps), complètement absents des débarquements déclarés, seraient également concernés. Il n'y a cependant jamais eu de marché intérieur pour ces derniers, et à part leur utilisation comme appâts, leur pêche est donc restée faible (Bouhlel 1988 ; Hosch 2010 ; Künzel et al. 1996a).5 A cause du manque d'information flagrant sur ces captures non-déclarées, nous avons repris donc l'allocation taxonomique utilisée pour la pêche artisanale déclarée.Enfin, une exploitation récente des holothuries est également présente à Djibouti (Hosch 2010). Ce phénomène est commun à de nombreux pays est-africains, en réponse à une très forte demande du marché asiatique (Hosch 2010 ; Le Manach et al. 2011). Les données publiées par la FAO ne contiennent pas d'holothuries ; nous avons donc utilisé les données fournies par la Direction de la Pêche (Tableau 1 ; pêche non-rapportée).• Rejets de la pêche artisanaleLes rejets de la pêche artisanale ne semblent pas importants à Djibouti, n'étant mentionnés dans aucune étude. Cependant, l'utilisation d'engins de pêche tels que le filet maillant entraîne souvent la capture d'espèces accessoires ou de petite taille (Kelleher 2005 ; Anon. 2010), généralement non débarquées. Il semble en effet peu probable que les pêcheurs gardent l'ensemble de ces captures comme appâts, et nous avons donc supposé que ces captures (estimée de manière tentative à 1% de la pêche artisanale totale (rapportées + non-rapportée) sur l'ensemble de la période)6 étaient rejetées. Étant donné les faibles tonnages, nous avons simplement alloué ces rejets au groupe générique 'poissons de fond'.• La pêche de subsistanceLa pêche de subsistance a toujours été considérée comme étant faible, bien que pouvant être importante à une échelle locale (Darar 1994 ; Morgan 2006 ; Hosch 2010). Elle a historiquement été pratiquée par quelques pêcheurs de Djibouti-ville et des autres régions côtières. Ces pêcheurs opèrent sur le plateau madréporique à l'aide de lignes à main ou de harpons, et ne possèdent généralement pas de bateaux (Morgan 2006 ; Hosch 2010). Aucune estimation de l'ensemble de la pêche de subsistance n'existe, mais ses caractéristiques se rapprochent d'un pays tel que la Mauritanie. En effet, à Djibouti comme en Mauritanie, il n'y a à l'origine pas de tradition de pêche, mais une tradition pastorale tournée vers la terre. La pêche de subsistance y a été estimée à 2% de la pêche artisanale déclarée (Belhabib et al. 2013), et il nous a semblé réaliste d'appliquer une valeur similaire pour Djibouti sur l'ensemble de la période étudiée.7 Ne possédant pas d'information quantitative, nous avons également alloué ces captures de subsistance au groupe générique 'poissons de fond'.5 Il y aurait eu quelques tentatives d'exportation à destination de l'Ethiopie, où le marché était plus ouvert (Morgan 2006). Morgan (2006) estimait également que 10 à 15% des débarquements étaient des petits pélagiques (sources et/ou méthodes de calcul non précisées), ce qui semble grandement surestimé étant donné les informations collectées sur l'ensemble de la période. 6 Cette proportion de rejets correspond à un taux de rejet moyen pour la petite pêche côtière des pays en voie de développement (Kelleher 2005 ; Anon. 2010).7 La Somalie est également un pays côtier à tradition pastoraliste, et beaucoup plus proche géographiquement que la Mauritanie. Cependant, la méthode employée par Persson et al. (ce volume) n'a pas pu être appliquée ici pour cause de manque de données concernant le nombre de bateaux.Tableau 1.  Captures des différentes espèces d'holothuries en t. Taxon 1996 1997 1998 1999 2000 2001 2002 2003 2004Holothuria scabra 15,0 18,0 17,0 16,0 15,0 16,0 12,0 5,0 2,0H. fuscogilva 0,0 0,0 5,0 6,0 7,0 8,5 6,0 5,0 1,0H. nobilis 0,0 0,0 0,2 0,3 0,3 0,5 0,3 0,1 0,1Actinopyga palauensis 2,0 5,0 10,0 12,0 14,0 12,0 13,0 11,0 7,0H. atra 5,0 12,0 25,0 24,5 24,0 26,0 25,0 16,0 15,0Total 22,0 35,0 57,2 58,8 60,3 63,0 56,3 37,1 25,1 18• La pêche récréativeLa pêche récréative semble encore peu développée et aucune étude concrète n'a encore été réalisée à ce sujet. En 2008–09, trois licences ont été octroyées à des entreprises de pêche sportive (Hosch 2010). Le système des licences fait également état de personnes possédant une licence de pêcheur non-commercial (i.e., pêchant de manière récréative principalement le week-end) au nombre de 60 en 2008 et 50 en 2009 (Hosch 2010). Enfin, il existe également des clubs de pêche au sein des bases militaires étrangères, comme l'ASAC Pêche de Djibouti pour l'armée française. Pour estimer les captures associées à ces trois types de pêches récréatives, nous avons procédé de plusieurs manières.Concernant la pêche sportive et les entreprises touristiques associées, nous avons pris contact avec la seule entreprise proposant des séjours de pêche à Djibouti pour les touristes étrangers ('Mémoire d'un Fleuve' ; www.memoiredunfleuve.com). Cette entreprise a  commencé son activité en 2001 et est active  20 semaines par an d'avril à fin juin, puis d'octobre à début novembre. La pratique de pêche est le no kill, c'est-à-dire qu'il y a remise à l'eau systématique des individus pêchés. Cependant, le responsable nous a indiqué que les guides sur place gardaient généralement un ou deux gros Scomberomorus commerson (thazards rayés) pêchés chaque semaine, soit environ 25 kg. Nous avons donc estimé que ce type de pêche récréative capturait 0,5 t de thazard par an (20 x 25 = 500 kg) depuis 2001. Il existe également des hôtels, structures de vacances qui proposent des activités pêche, mais nous ne possédions aucun renseignement précis sur ces acteurs. Nous avons donc considéré que les 500 kg de thazards pêchés annuellement représentaient la moitié des captures de 'Mémoire d'un Fleuve' (l'autre moitié étant allouée au groupe générique 'poissons pélagiques'), et que les autres sources de pêche récréative représentait deux fois ces captures (toutes attribuées au groupe 'poissons pélagiques').Concernant les pêcheurs récréatifs locaux pratiquant la pêche le week-end, cette pêche est ancienne de par la présence française pré- et post-indépendance, et sa pratique est concentrée à Djibouti. El Gharbi (1987) mentionnait 612 pêcheurs 'amateurs' autorisés à pêcher en 1987. Cependant, ces chiffrens semblent être très sur-estimés, étant donné que la distinction entre pêcheurs récréatifs et pêcheurs artisans est difficile à faire à cause du faible coût de l'autorisation et la pratique de la pêche artisanale à mi-temps. En 2008, nous avons donc considéré qu'il y avait 55 pêcheurs récréatifs en 2010, suivant les chiffres proposés par la Direction de la Pêche (60 pêcheurs en 2008 et 50 en 2009). Nous avons ensuite estimé le nombre de pêcheurs récréatifs annuels au pro rata du nombre d'habitants à Djibouti-ville (Guillaume 1979 ; République de Djibouti).8 Finalement, nous avons considéré que ces pêcheurs capturaient cinq kilos de 'poissons pélagiques' par sortie, 24 fois par an (i.e., deux week-ends par mois), soit 120 kg·pêcheur-1·année-1. Enfin, concernant les clubs de pêche au sein des armées, comme le club ASAC de l'armée française enregistré auprès des Clubs Sportifs et Artistiques de la Défense (CSAD).9 Ces pêcheurs pratiquent également majoritairement la remise à l'eau, mais les guides locaux gardent une partie des poissons lors des sorties. Ainsi, en 2010, 20% des captures d'un voyage de pêche sont allées aux guides (www.youtube.com/watch?v=TAjsjWg6o7o), soit 0,5 t de 'poissons pélagiques' (V. Cressy, comm. pers.). Nous avons donc inclus les clubs de pêche des armées étrangères dans notre reconstruction en estimant la capture à 0,5 t par an depuis 2000, n'ayant pas plus de détails sur les effectifs historiques de l'ensemble des clubs et les pratiques associées. Reconstruction des captures djiboutiennes à l'extérieur de la ZEE nationaleLes captures djiboutiennes à l'extérieur de la ZEE nationale sont supposées assez rares, étant donné le bon état des stocks et la petite échelle des activités (Hosch 2010). Cependant, au début des années 80, certains pêcheurs étaient connus pour fréquenter les eaux somaliennes. El Gharbi (1987) estimait que 42% des captures débarquées à l'ACPM en 1986 provenaient de Somalie, soit 30,1% des captures totales de la pêche artisanale déclarée. En 1987, une interdiction de l'accès aux zones somaliennes a été prononcée et a possiblement stoppé le phénomène (El Gharbi 1987) pour quelques temps. Pendant le conflit des années 90, les activités de pêche se sont cependant concentrées dans le sud du pays, engendrant une nouvelle expansion des activités dans la ZEE somalienne (Künzel et al. 1996a). Enfin, une pêche dans les eaux somaliennes est pratiquée depuis 2008 par trois boutres (>16 m) possédées par l'entreprise 'Red Sea Fishing' (RSF). Ces trois boutres ciblent les démersaux nobles avec des nasses et parfois les crevettes au chalut. Ces techniques sont très différentes de celles pratiquées par les autres pêcheurs djiboutiens (Hosch 2010). Les informations fournies par la Direction de la Pêche et l'entreprise RSF font état de 50% des captures de RSF provenant de Somalie pour l'année 2010, avec une répartition par espèce similaire au reste des débarquements. En 2010, RSF représentait 9,12% des captures nationales déclarées. Il existe également un armateur privé, 'Pêcherie de Loyada', qui opère quelques bateaux pêchant presque uniquement dans les eaux somaliennes et débarquant à la frontière.A partir des données présentées ci-dessus, nous avons estimé les captures faites hors de la ZEE djiboutienne en construisant une série temporelle des captures artisanales provenant des eaux somaliennes (seul pays mentionné pour ces pratiques) à partir des données de pêche artisanale déclarées dans les eaux djiboutiennes. Pour la période pré-indépendance, nous avons considéré que la pêche dans les eaux somaliennes était nulle à cause des faibles moyens techniques et du bon état des ressources halieutiques. Le pourcentage a ensuite augmenté linéairement à partir de 1980 pour atteindre 30,1% en 1986 suite au développement de la pêche artisanale et un accroissement de la fréquentation de ces zones. Nous avons ensuite estimé que cette valeur a été nulle entre 1987 et 2007 suite à 8 Ces deux publications nous ont permis de recréer une série temporelle du nombre d'habitants à Djibouti-ville. Une interpolation linéaire a été faite entre les différents points d'ancrage.9 Ce groupe existe depuis une dizaine d'année et compte entre 10 et 25 membres selon les années (Vincent Cressy, ex-trésorier ASAC, comm. pers.).Fisheries catch reconstruction for Djibouti — Colléter et al. 19l'interdiction de fréquenter les eaux somaliennes, sauf entre 1992 et 1994, où elle a été égale à la moitié de la valeur de 1986 (15,05%). En effet, il nous a semblé raisonnable de considérer une valeur moitié moins importante afin de refléter l'impact du conflit sur l'étalement des zones de pêche. Enfin, pour les années 2008–10, nous avons estimé que le pourcentage des captures hors Djibouti était égal à la moitié du ratio des captures artisanales déclarées par RSF, soit 4,56%. La répartition des captures par famille/espèce pour ces captures est la même que pour la pêche artisanale déclarée dans les eaux djiboutiennes.Reconstruction des captures étrangères dans la ZEE djiboutienneLa pêche étrangère dans les eaux djiboutiennes est une activité non prise en compte dans les captures totales. Ceci est préjudiciable, car il apparaît que ce phénomène est très important et ancien (Morgan 2006). Des bateaux étrangers (principalement originaires du Yémen, mais aussi de Somalie) y ont pêché depuis les années 50, notamment le long de la côté Nord, où les pêcheurs yéménites "hantaient la côte Nord […] aux termes d'accords de péage mystérieux" (Pujo 1967). Ces pêcheurs sont mentionnés dans de nombreux travaux (e.g., Allain 1974 ; Künzel et al. 1996a ; Morgan 2006), et ont constitué une concurrence mieux organisée, débarquant parfois à Djibouti et submergeant ainsi le marché local (Clouet 1970). Après l'indépendance, ces pratiques ont continué, les capacités de contrôle en mer restant limitées (Morgan 2006). A ce jour, des centaines d'embarcations généralement bien plus importantes que celles utilisées par les djiboutiens continuent d'y prendre part, avec des captures représentant entre 30–35% (Direction de la Pêche, données non publiées) et 50% (Hosch 2010) des captures totales officielles (majoritairement yéménites). Nous avons donc estimé une pêche artisanale étrangère représentant 42,5% des captures totales de la pêche artisanale déclarée sur toute la période. Les flottilles djiboutiennes et étrangères n'ayant pas connu la même vitesse de développement, nous avons utilisé les données de débarquements de la pêche artisanale somalienne (Persson et al., ce volume) et yéménite (Tesfamichael et al., 2012) afin d'ajuster l'allocation des captures à ces deux pays pour l'ensemble de la période étudiée. Nous avons considéré une pêche yéménite représentant 35% sur les 42,5% cités plus haut pour l'année 2010, et estimé les autres années au pro rata de l'évolution des débarquements artisanaux calculés par Tesfamichael et al. (2012) et Persson et al., (ce volume; au final, ces proportions ont varié de 28,8 à 39,3% de pêche somalienne, et le reste de pêche yéménite). Bien qu'il existe quelques informations sur les espèces ciblés,10 nous avons gardé ici encore la composition taxonomique de la pêche artisanale.résultats et disCussionReconstruction des captures domestiques dans la ZEE nationalePour la période pré-indépendance, les captures reconstruites sont bien plus faibles que celles publiées par la FAO.11 Ces données reconstruites offrent une lecture plus juste de l'évolution réelle des pêches à Djibouti, avec une augmentation d'environ 130 à près de 300 t entre 1950 et 1978, liée à l'augmentation 10 Par exemple, concernant le Yémen, les sardinelles étaient historiquement ciblées dans le Nord de Djibouti par ces opérations (Clouet 1970 ; Allain 1974). Les requins ont également été historiquement ciblés, notamment par la tribu Hakimé durant des campagnes de un ou deux mois (Clouet 1970). Il semblerait que les populations de requins soient maintenant sur-exploitées à cause de cette pêche étrangère, qui s'est par ailleurs intensifiée (Künzel et al. 1996a). Ces opérations de pêche cibleraient également de nombreuses autres espèces démersales et pélagiques (Darar et Hosch 2010).11 Les données publiées par la FAO (déclarées par la France) restent inexpliquées, nous n’avons pu en trouver ni la source, ni des explications associées. -      0,5    1,0    1,5    2,0    2,5  Subsistance et récréative Artisanale Captures rapportées à la FAOAutresScombridae Sphyraena Lethrinidae Carangidae Scomberomorus spp. Lutjanidae  -      0,5    1,0    1,5    2,0    2,5  Captures totales (milliers de t)Année1950 1960 1970 1980 1990 2000 2010 ABSerranidaeFigure 2.  Captures djiboutiennes de 1950 à 2010 dans la ZEE nationale, A) par secteurs et comparées aux données déclarées à la FAO, et B) par taxons pêchés. Voir Tableau Annexe A1 et Tableau Annexe A2 pour plus de détails. 20du nombre de pêcheurs (Figure 2). Pour la période 1950–1967, l'hypothèse forte d'une CPUE constante mais plus faible que pour les années suivantes est justifiée par le progrès technique substantiel entre les deux périodes (e.g., amélioration des bateaux, développement du monofilament) et une dynamique plus active de la filière (El Gharbi 1987 ; Künzel et al. 1996a). La reconstruction de la période post-indépendance a également permis de mieux comprendre les données FAO et de mieux refléter l'évolution des pêches à Djibouti. Les données FAO étaient différentes de celles de la Direction de la Pêche jusqu'en 2005, année à partir de laquelle plus aucune correction n'a été appliquée. Les corrections appliquées entre 1983 et 1991 (i.e., 70% des captures déclarées ; FAO 1991) ne correspondaient pas à 70% des données que la Direction de la Pêche nous a transmises.12 Les données reconstruites permettent de visualiser les différentes phases de l'évolution de la pêche post-indépendance (Figure 2). Les captures ont rapidement augmenté jusqu'à plus de 1 000 t à la fin des années 80, ce qui a marqué le début d'un engagement réel des pouvoirs publics en faveur du développement de la pêche artisanale avec une amélioration de la qualité de l'intervention et de l'appui des pouvoirs publics. Les captures ont ensuite diminué jusqu'en 1995, ce qui s'explique par les difficultés financières des gérants de la Pêcherie de Boulaos, mais aussi la guerre civile au Nord du pays de 1992 à 1994. Elles ont ensuite augmenté de nouveau jusqu'en 2004 pour atteindre plus de 2 000 t. Durant cette période, les activités de pêche se sont re-développées et les infrastructures de débarquement de la Pêcherie de Boulaos ont été reprises par un opérateur privé. Après une nouvelle baisse, ce niveau a de nouveau été atteint en 2010. Cette dernière période a été caractérisée par (i) l'ouverture des infrastructures de débarquement du port de pêche de Djibouti-ville, financée par la Banque Africaine de Développement (BAD), (ii) la réhabilitation des sites de débarquement d'Obock et de Tadjourah par la Coopération française, et (iii) l'arrivée de nouveaux opérateurs privés. Cette volonté de développement s'est surtout traduite par la concentration des moyens concernant le port de pêche de Djibouti-ville, plus facile à contrôler. Au final, les captures de la pêche artisanale ont été démultipliées entre 1950 et 2010, sans aucune pêche industrielle. Le développement de la pêche à Djibouti a donc exclusivement focalisé sur les pêcheries artisanales, soutenu par une volonté gouvernementale en plusieurs phases au cours des soixante dernières années.Nous avons également complètement amélioré la composition taxonomique de ces captures (Figure 2B). Cette nouvelle allocation repose sur les taxons déclarés à la FAO dans le courant des années 80 et met en évidence l'importance des Serranidae (17%), des Lutjanidae (15%), des Scomberomorus spp. (13%), des Carangidae (11%) et des Lethrinidae (9%). Étant donné le peu d'informations disponibles sur cette composition taxonomique, celle-ci nous paraît être une amélioration substantielle par rapport aux données publiées par la FAO, en ce qui concerne la période pré-indépendance. Un travail de reconstruction plus poussé pourrait cependant être envisagé afin de refléter les changements temporels qui ont eu lieu au niveau des espèces pêchées.Notre reconstruction met également en avant une part importante non déclarée des captures djiboutiennes. Ceci est d'autant plus vrai dans le contexte de faiblesse de moyens du service de la Direction de la Pêche et l'absence d'enquête sur la consommation de poisson à mettre en regard avec les données de production. Il est ainsi difficile de faire des recoupements d'informations, et il existe toujours par exemple des circuits de vente directe aux restaurateurs par des personnes pour qui la pêche est une activité partielle. Pour la période pré-indépendance, ce phénomène était dû majoritairement à une filière encore peu développée et une commercialisation par des circuits non contrôlés. Plus récemment, ce phénomène a principalement été dû aux débarquements faits par les pêcheurs djiboutiens de la côte Nord au Yémen où les avantages économiques sont nombreux (Hosch 2010). L'ensemble de ces hypothèses montre l'importance d'améliorer les moyens de suivi et contrôle de la Direction de la Pêche afin de mieux connaître l'importance de ces captures et leur composition spécifique. Même si les secteurs de subsistance et de pêche récréative sont marginaux en terme de tonnages, ce constat s'y applique également. Ceci met en avant la nécessité d'augmenter aujourd'hui les moyens alloués à la Direction de la Pêche afin de mieux contrôler et évaluer ces pratiques encore très peu étudiées. De plus, plusieurs indices peuvent indiquer la surexploitation de certains stocks (détaillés plus bas). Il semble donc nécessaire de mettre en place un processus de récolte de données exhaustif quant aux différents secteurs contribuant aux captures dans les eaux djiboutiennes. Ceci est nécessaire afin de pouvoir réaliser des diagnostics et avis scientifiques fiables pour une exploitation durable des ressources.Reconstruction des captures djiboutiennes en dehors de la ZEE domestique, et pêches étrangèresNotre estimation des captures hors ZEE repose sur des hypothèses fortes qu'il conviendrait de préciser par l'inclusion de nouvelles données et une analyse détaillée de ce phénomène. Cette série temporelle représente une première approche caricaturale du phénomène décrit (Figure 3), et nous avons utilisé les seules mentions faites d'une pêche djiboutienne opérée en dehors de la ZEE sur l'ensemble de la période. A l'heure actuelle, les seules opérations recensées concernent l'entreprise privée RSF qui pratique des techniques différentes (i.e., nasses, et chalut interdit dans la ZEE djiboutienne) et ceux sans accord particulier signé avec la Somalie. Nous avons supposé que la répartition par famille/espèce était la même que la pêche artisanale déclarée effectuée dans la ZEE djiboutienne, mais ceci reste à consolider. Il conviendrait de mieux encadrer et estimer ces pratiques aujourd'hui assez faibles afin de pouvoir établir des diagnostics justes, et peut-être mettre en place des accords précis avec la Somalie qui pêche également dans les eaux djiboutiennes. Ces captures somaliennes mais surtout yéménites dans la ZEE djiboutienne équivalent à des captures considérables, puisqu'elles ont atteint plus de 1 500 t en 2010 (Figure 3). Bien que nous ayons gardé un ratio pêche étrangère:pêche domestique constant au cours du temps, il est cependant possible que ce ratio ait été plus important pendant la période pré-indépendance à cause du faible développement de la pêche djiboutienne en 12 De 1983 à 1985, les données FAO correspondaient à 70% des seules captures de l'ACPM, puis à 100% en 1986 (El Gharbi 1987). Enfin, les corrections appliquées entre 1992 et 2004 n'étaient, à notre connaissance, expliquées nulle part.Fisheries catch reconstruction for Djibouti — Colléter et al. 21comparaison de celle en provenance de la Somalie et du Yémen. Cette pêche est majoritairement pratiquée par les Yéménites dont la présence est attestée depuis longtemps (Clouet 1970 ; Hosch 2010 ; Morgan 2006). Les capacités de patrouille en mer étant limitées et les fonds riches, il est logique qu'une telle activité soit apparue de par l'importance de la flotte yéménite sur la période 1950–2010 et leur tradition de pêche. La limite entre opérations djiboutiennes et étrangères dans la ZEE est cependant parfois floue. En effet, la présence yéménite ancienne se traduit aujourd'hui par des contournements, comme celui de l'obligation d'être citoyen djiboutien pour obtenir une licence. Il existe apparemment des pêcheurs yéménites en possession d'une licence djiboutienne (et d'un bateau immatriculé à Djibouti) leur permettant de pêcher dans la ZEE, ces captures n'étant ensuite pas débarquées à Djibouti. Toutes sortes d'opérations de ce genre ont été reportées. Dans cette étude, nous nous sommes basés exclusivement sur le pavillon et la zone, mais il faut garder à l'esprit que cela occulte une série de comportements difficiles à catégoriser. Par manque d'information, nous avons utilisé la composition taxonomique des captures djiboutiennes. Il existe cependant quelques informations laissant penser que les captures yéménites laissent une plus grande part aux requins (e.g., Carcharhinus brevipinna et autres Carcharhinidae) qui sont préférentiellement ciblés, ainsi qu'aux petits pélagiques dont la consommation est plus répandue au Yémen (Abbes 1985 ; Clouet 1970 ; Hosch 2010). Devant l'importance de ce secteur et les captures illicites engendrées, il conviendrait de mieux contrôler ces opérations afin d'en connaître l'étendu et ainsi garantir une gestion durable de la pêche artisanale djiboutienne. Plusieurs éléments semblent en effet indiquer la surexploitation de certains stocks : les ressources en holothuries auraient déjà été exploitées jusqu'à épuisement commercial, et les pêcheurs notent que certaines strates bathymétriques sont surexploitées (les vivaneaux et les mérous se seraient apparemment raréfiés entre 30 et 50 m), ou que les migrations saisonnières des grands pélagiques se font plus irrégulièrement et moins fortement qu'avant (Hosch 2010). De plus, les pêcheurs se plaignent d'une abondance accrue de dauphins responsables d'une déprédation importante de leurs prises, phénomène pouvant être lié à l'exploitation intense des requins dans la ZEE djiboutienne (Hosch 2010). Nous avons estimé que la capture totale de requins dans la ZEE djiboutienne s'élevait à environ 116 t en 2010, ce qui est supérieur au potentiel estimé de capture de 70 t par année (Darar 1994).ConClusionCette étude nous a permis de reconstruire l'ensemble des captures par la pêche à Djibouti de 1950 à 2010. Nous avons ainsi pu mieux comprendre, réviser et compléter les données de la série FAO. Les données reconstruites comprennent notamment l'évolution de la pêche artisanale djiboutienne : la capture est restée très limitée durant la période pré-indépendance avec l'ancrage d'une tradition pastorale dans les coutumes. La période post-indépendance a ensuite vu le développement d'une flottille artisanale professionnelle de par la volonté des pouvoirs publics. De plus, la prise en compte de plusieurs secteurs tels que la pêche artisanale non-déclarée, la pêche récréative, et la pêche de subsistance affine le diagnostic sur l'évolution du secteur de la pêche. Il existe aujourd'hui peu de données et d'enquêtes précises sur ces secteurs qui, bien qu'ayant des tonnages faibles, peuvent avoir un impact substantiel sur les ressources. Il semble donc essentiel d'augmenter les moyens alloués à la Direction de la Pêche pour l'encadrement et le contrôle des différentes activités de pêche afin d'obtenir de meilleures estimations, et ainsi garantir une exploitation durable des ressources. L'ajout des captures illégales opérées par des pêcheurs étrangers illégaux (du Yémen majoritairement) renforce ce diagnostic puisqu'ils prélèvent sans contrôle et de manière importante des ressources halieutiques. Au final, il semble aujourd'hui nécessaire d'établir des diagnostics plus précis de l'impact de la pêche sur les stocks ciblés. En effet, plusieurs éléments pourraient indiquer une surexploitation de certains stocks ciblés, ce qui pourrait avoir des répercussions néfastes sur les écosystèmes marins djiboutiens.remerCiementsCette étude a été réalisée dans le cadre du projet Sea Around Us, une collaboration entre l'Université de Colombie Britannique et le Pew Charitable Trusts. MC, FLM et DP remercient également la Paul G. Allen Family Foundation pour son soutien financier.Captures djiboutiennesen dehors de la ZEE -      0,5    1,0    1,5    2,0  Captures totales (milliers de t)Année1950 1960 1970 1980 1990 2000 2010 SomalieYémenFigure 3.  Captures étrangères dans la ZEE djiboutienne et captures djiboutiennes à l'extérieur de la ZEE nationale, 1950–2010.(voir Tableau Annexe A1 pour détails). 22réFérenCesAbbes R (1985) Bilan des connaissances acquises sur la faune et la flore sous-marines de la région de Djibouti. Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER), Nantes (France). 6 + xxxiv p.Allain C (1974) Mission d'information sur le développement de la pêche dans le Territoire des Afars et des Issas — du 10 au 15 avril 1974. Office Scientifique et Technique des Pêches Maritimes (ISTPM), Nantes (France). i + 8 p.Anon. (2010) The hidden harvests — The global contribution of capture fisheries. Agriculture and Rural Development Department, Sustainable Development Network. Prepared by The World Bank, the Food and Agriculture Organization of the United Nations (FAO), and the WorldFish Center, Washington, DC (USA). 99 p.Anon. (2011) Pour un plan stratégique de recherche sur les ressources et l'environnement marins à Djibouti — Compte-rendu de l'atelier sur les ressources et l'environnement marins. Université de Djibouti, Faculté des Sciences, Djibouti. 17 p.Augustin P (1971) Quelques dénombrements de la population du Territoire français des Afars et des Issas (suite). POUNT (Bulletin de la Société d'Etudes de l'Afrique Orientale) 4ème Année(14): 15–38.Belhabib D, Gascuel D, Kane EA, Harper S et Zeller D (2013) Preliminary estimation of realistic fisheries removals from Mauritania, 1950–2010. pp. 63–80, Dans: Belhabib D, Zeller D, Harper S and Pauly D (éds.), Marine fisheries catches in West Africa, 1950–2010, part I. Fisheries Centre Research Reports 20 (3). University of British Columbia, Vancouver (Canada).Bjoerklund I et Walter-Dehnert G (1983) Baseline socio-economic and marketing study of artisanal fisheries in the Republic of Djibouti. Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie) 80 p.Bouhlel M (1988) Poissons de Djibouti. Dubai Printing Press. 416 p.CITES (2002) Interprétation et application de la Convention — Commerce et conservation des espèces — Commerce des concombres de mer des familles Holothuridae et Stichopodidae. Douzième session de la Conférence des Parties (CdP 12) 45, Convention sur le commerce international des espèces de faune et de flore sauvages menacées d'extinction (CITES), Santiago (Chili). 29 p.CITES (2006) Concombres de mer. Vingt-deuxième session du Comité pour les animaux (AC 22) 16, Convention sur le commerce international des espèces de faune et de flore sauvages menacées d'extinction (CITES), Lima (Pérou). 32 p.Clouet A (1970) La pêche à Djibouti. La Revue Maritime 275: 476–493.Darar A (1994) An account of fisheries development in the Republic of Djibouti with notes on the growth and mortality of three species of groupers. Naga, The ICLARM Quarterly 17(2): 30–32.Darar A et Hosch G (2010) Analyse des implications pour la mise en opération d'embarcations de pêche industrielles à Djibouti. Direction des Pêches de Djibouti, Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie). i + 9 p.Devinat P (1957) La mer Rouge et les intérêts français. Politique Etrangère 22ème année(3): 311–324.El Gharbi R (1987) La pêche artisanale dans la République de Djibouti — Analyse bio-économique. Projet de développement de la pêche artisanale, Phase II, United States Agency for International Development (USAID), Resource Development Associates International (RDA), Djibouti. 59 + xxix p.Emerton L (1998) Djibouti biodiversity: economic assessment. Djibouti national biodiversity strategy and action plan, International Union for Conservation of Nature (IUCN), Djibouti. 54 p.FAO (1991) Fishery statistics — Catches and landings. FAO yearbook 72, Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie).FAO (2012) FishStat Plus — Universal software for fishery statistical time series. v2.3. Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie).Guillaume A (1979) Djibouti. pp. 41–50 In L'évaluation des effectifs de la population des pays africains — Tome 1. Groupe de Démographie Africaine IDP-INED-INSEE-MINCOOP-ORSTOM, Paris (France).Harper S and Zeller D, editors (2012) Fisheries catch reconstructions: islands, part II. Fisheries Centre Research Reports 19 (4). University of British Columbia, Vancouver (Canada). 143 p.Harper S, Zylich K, Boonzaier L, Le Manach F, Pauly D and Zeller D, editors (2012) Fisheries catch reconstructions: islands, part III. Fisheries Centre Research Reports 20 (5). University of British Columbia, Vancouver (Canada). 134 p.Hosch G (2010) Plan d'action national visant à prévenir, à contrecarrer et à éliminer la pêche illicite, non déclarée et non réglementée — République de Djibouti. Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie). v + 54 p.IFAD (1984) Rapport de la mission de formulation de renforcement du projet pilote de développement de la pêche artisanale. Projet de développement de la pêche artisanale, Phase I, International Fund for Agricultural Development (IFAD), Djibouti. 1 p.Kelleher K (2005) Discards in the world's marine fisheries — An update. FAO Fisheries Technical Paper 470, Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie). 131 p.Künzel T, Darar A et Vakily JM (1996a) Composition, biomasses et possibilités d'exploitation des ressources halieutiques djiboutiennes — Tome 1 — Analyse. Ministère de l'Agriculture et de l'Hydraulique, Direction de l'Elevage et des Pêches, République de Djibouti, Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ). ix + 63 p.Künzel T, Darar A et Vakily JM (1996b) Composition, biomasses et possibilités d'exploitation des ressources halieutiques djiboutiennes — Tome 2 — Données. Ministère de l'Agriculture et de l'Hydraulique, Direction de Fisheries catch reconstruction for Djibouti — Colléter et al. 23l'Elevage et des Pêches, République de Djibouti, Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) GmbH, Bundesrepublik Deutschland, Djibouti. 156 p.Le Manach F, Gough C, Humber F et Harper S (2011) Reconstruction of total marine fisheries catches for Madagascar. pp. 21–37 In Harper S et Zeller D (eds.), Fisheries catch reconstructions: Islands, Part II. Fisheries Centre Research Reports 19 (4). University of British Columbia, Vancouver (Canada).Lewin WC, Arlinghaus R et Mehner T (2006) Documented and potential biological impacts of recreational fishing: insights for management and conservation. Reviews in Fisheries Science 14(4): 305–367.Moal RA (1969) Les perspectives de développement de Djibouti, port de pêche industrielle. POUNT (Bulletin de la Société d'Etudes de l'Afrique Orientale) 2ème Année(7): 39–42.Moal RA et Grateau J (1967) Pêche en Territoire Français des Afars et des Issas. POUNT (Bulletin de la Société d'Etudes de l'Afrique Orientale) 1ère Année(3): 17–24.Morgan G (2006) Country review: Djibouti. pp. 195–201 Dans De Young C (éd.) Review of the state of world marine capture fisheries management: Indian Ocean. FAO Fisheries and Aquaculture Technical Papers 488. Organisation des Nations Unies pour l'Alimentation et l'Agriculture (FAO), Rome (Italie).Pauly D (1998) Rationale for reconstructing catch time series. EC Fisheries Cooperation Bulletin 11(2): 4–10.Persson L, Lindop A, Harper S, Zylich K and Zeller D (this volume) Failed state: reconstruction of domestic fisheries catches in Somalia 1950–2010.Pujo JM (1967) Les boutres à Djibouti : une survivance de l'âge de la voile. POUNT (Bulletin de la Société d'Etudes de l'Afrique Orientale) 1ère Année(2): 9–16.République de Djibouti (2012) Annuaire statistique de Djibouti — Edition 2012. Ministère de l'économie et des finances chargé de l'industrie et de la planification. 121 p.Rouaud A (1997) Pour une histoire des Arabes de Djibouti, 1896–1977. Cahiers d'Etudes Africaines 37(146): 319–348.Tesfamichael D, Rossing P and Saeed H (2012) The marine fisheries of Yemen with emphasis on the Red Sea and cooperatives. Pp. 105–152. In Tesfamichael D and Pauly D (eds.) Catch reconstruction for the Red Sea large marine ecosystem by countries (1950–2010). Fisheries Centre Research Reports 20(1). Fisheries Centre, University of British Columbia [ISSN 1198–6727].Waldstein AS et Lampe H (1988) Djibouti fisheries development project. AID evaluation working paper 110, United States Agency for International Development (USAID), Djibouti. xiv + 45 p.Zeller D and Harper S, editors (2009) Fisheries catch reconstructions: islands, part I. Fisheries Centre Research Reports 17 (5). University of British Columbia, Vancouver (Canada). 108 p.Zeller D et Pauly D, editors (2007) Reconstruction of marine fisheries catches for key countries and regions (1950–2005). Fisheries Centre Research Reports 15 (2). University of British Columbia, Vancouver (Canada). 163 p. 24Tableau Annexe A1.  Captures domestiques reconstruites (par secteur) et rapportées à la FAO, et captures étrangères et en dehors de la ZEE nationale, de 1950 à 2010.Années Captures domestiques Captures étrangères dans la ZEE nationaleDans la ZEE nationale Hors ZEEArtisanales Récréatives Subsistance Total reconstruit Total rapporté à la FAO Somalie Yémen1950  127  1  2  130  500 -  28  73 1951  131  1  2  134  800 -  29  75 1952  134  1  2  137  500 -  27  79 1953  137  1  2  140  500 -  27  82 1954  140  1  2  144  900 -  28  83 1955  144  1  2  147  800 -  26  88 1956  147  1  2  150  500 -  35  81 1957  150  1  2  154  600 -  35  84 1958  154  1  2  157  800 -  37  85 1959  157  1  2  160  600 -  35  89 1960  160  1  3  164  900 -  38  89 1961  163  1  3  167  800 -  39  90 1962  167  1  3  171  900 -  40  92 1963  170  1  3  174  1 000 -  41  93 1964  173  1  3  177  300 -  40  98 1965  177  2  3  181  300 -  41  99 1966  180  2  3  184  300 -  42  100 1967  183  2  3  188  300 -  42  103 1968  192  2  3  197  300 -  45  107 1969  202  2  3  207  300 -  48  111 1970  211  2  3  216  300 -  51  116 1971  220  2  3  226  300 -  56  118 1972  229  2  4  235  100 -  55  127 1973  239  2  4  245  200 -  49  140 1974  248  3  4  254  380 -  47  150 1975  257  3  4  264  300 -  49  154 1976  266  3  4  273  230 -  45  166 1977  275  3  4  283  230 -  47  171 1978  253  3  4  260  230 -  45  155 1979  316  3  5  325  231 -  53  197 1980  392  3  6  402  251 13  59  252 1981  486  3  8  497  385 33  76  309 1982  505  3  8  516  425 52  90  310 1983  581  4  9  594  409 79  114  346 1984  606  4  10  619  409 103  116  364 1985  529  4  8  541  380 108  95  324 1986  789  4  13  805  409 188  152  473 1987  843  4  13  861  426 -  148  520 1988  971  4  15  990  454 -  160  609 1989  807  4  13  824  399 -  119  520 1990  1 048  4  17  1 069  361 -  125  705 1991  894  4  14  912  253 -  90  618 1992  436  5  7  447  276 52  40  305 1993  260  5  4  269  301 31  21  185 1994  442  5  7  454  321 53  33  317 1995  505  5  8  518  351 -  30  370 1996  590  5  9  604  401 -  51  399 1997  962  5  15  981  501 -  87  647 1998  1 017  5  15  1 037  601 -  103  657 1999  1 066  5  16  1 088  701 -  124  674 2000  1 021  6  15  1 042  801 -  150  611 2001  1 081  8  16  1 105  901 -  149  657 2002  1 117  8  17  1 142  1 001 -  135  705 2003  1 074  8  16  1 099  1 101 -  109  712 2004  1 099  8  17  1 124  1 201 -  110  741 2005  1 983  9  31  2 023  1 571 -  229  1 342 2006  1 641  9  26  1 676  1 299 -  192  1 108 2007  1 552  9  25  1 585  1 229 -  228  1 001 2008  1 518  9  24  1 550  1 206 55  297  905 2009  1 270  9  20  1 299  1 058 46  207  799 2010  2 007  9  32  2 048  1 058 73  281  1 309 Fisheries catch reconstruction for Djibouti — Colléter et al. 25Tableau Annexe A2.  Composition taxonomique des captures domestiques dans la ZEE djiboutienne de 1950 à 2010.Années Epinephelus spp. Lutjanidae Scomberomorus spp. Carangidae Lethrinidae Sphyraena Scombridae Mugilidae Autres1950  15  19  25  14  14  5  10  3  26 1951  16  19  25  14  14  6  10  3  27 1952  16  20  26  15  14  6  10  3  27 1953  16  20  26  15  15  6  10  3  28 1954  17  21  27  15  15  6  11  3  29 1955  17  21  28  16  15  6  11  3  29 1956  18  22  28  16  16  6  11  3  30 1957  18  22  29  17  16  6  11  3  31 1958  18  23  30  17  16  7  12  3  31 1959  19  23  30  17  17  7  12  3  32 1960  19  24  31  18  17  7  12  3  33 1961  20  24  31  18  18  7  13  3  33 1962  20  25  32  18  18  7  13  3  34 1963  20  25  33  19  18  7  13  3  35 1964  21  26  33  19  19  7  13  4  36 1965  21  26  34  19  19  8  14  4  36 1966  22  27  35  20  19  8  14  4  37 1967  22  27  35  20  20  8  14  4  38 1968  23  29  37  21  21  8  15  4  40 1969  24  30  39  22  22  9  15  4  42 1970  25  31  41  23  23  9  16  4  44 1971  26  33  42  24  24  9  17  4  46 1972  28  34  44  25  25  10  18  5  48 1973  29  35  46  26  26  10  18  5  50 1974  30  37  48  27  27  11  19  5  52 1975  31  38  50  28  28  11  20  5  53 1976  32  40  51  29  29  11  20  5  55 1977  33  41  53  30  30  12  21  6  57 1978  30  38  49  28  27  11  19  5  53 1979  38  47  61  35  34  13  24  6  66 1980  47  58  76  43  42  17  30  8  81 1981  58  72  94  54  52  21  37  10  99 1982  61  75  97  56  54  22  39  10  103 1983  4  105  45  96  72  31  66  41  134 1984  69  54  107  103  48  48  66  12  112 1985  55  74  105  66  70  21  47  4  99 1986  157  160  210  65  78  15  44  6  71 1987  118  104  178  39  90  27  33  6  265 1988  214  125  55  112  91  32  68  15  279 1989  190  286  74  64  74  18  22  2  94 1990  239  348  164  60  83  29  58  1  87 1991  270  193  168  46  98  42  39  1  57 1992  131  94  81  22  48  20  19  0  31 1993  79  56  49  13  29  12  11  0  20 1994  134  94  83  22  49  20  19  0  32 1995  143  114  96  29  57  26  21  0  33 1996  140  112  91  28  56  28  21  0  127 1997  220  165  112  64  73  64  26  0  257 1998  221  158  95  79  71  79  24  0  309 1999  228  157  85  100  71  93  21  0  333 2000  214  148  71  107  71  95  30  0  305 2001  224  157  67  123  78  106  39  0  310 2002  236  157  63  131  84  121  47  0  302 2003  229  149  56  131  84  121  56  0  273 2004  235  150  53  137  89  128  66  0  265 2005  361  221  53  216  136  204  115 -  717 2006  244  138  374  204  158  185  134 -  240 2007  228  123  339  263  174  145  161 -  154 2008  157  214  123  257  140  212  240 -  207 2009  158  213  97  242  150  143  119 -  177 2010  250  336  154  383  237  225  188 -  275 Fisheries catch reconstructions for the îles éparses — Le Manach and Pauly 27First estimate oF unreported CatCh in the FrenCh Îles éparses, 1950-2010*Frédéric Le Manach1,2† and Daniel Pauly11 Sea Around Us, Fisheries Centre, University of British Columbia, 2202 Main Mall, Vancouver V6T 1Z4, Canada2 Institut de Recherche pour le Développement, UMR212 Ecosystèmes Marins Exploités, Avenue Jean Monnet, CS 30171, 34203 Sète cedex, France† Current address: BLOOM Association, 77 rue du Faubourg Saint-Denis, 75010 Paris, Francefredericlemanach@bloomassociation.org; d.pauly@fisheries.ubc.caabstraCtIn this report, we used the catch reconstruction approach developed by the Sea Around Us to estimate the total marine fisheries catch in the EEZs of the Îles Éparses. These islands being uninhabited, there are no records of such fisheries in the official fisheries data published by the Food and Agriculture Organization of the United Nations (FAO), and thus, our reconstruction is entirely comprised of unreported catches. Catches were estimated to around 2,800 tonnes between 1989 and 2010, essentially in the Glorieuses Archipelago. The small-scale artisanal barques from Mayotte represented 76.8% of the total, followed by the recreational and semi-industrial handline fisheries, with 14.0% and 6.7%, respectively. Lutjanus bohar represented 39.9% of the catch, followed by Serranidae, yellowfin tuna, other Scombridae and Carangidae, representing 21.5%, 7.0%, 7.3%, and 5.5% of the catch, respectively (the rest being composed of various species  of groundfishes and pelagic fishes).introduCtionThe Îles Éparses (i.e., 'Scattered Islands') encompass a group of five small entities dispersed around Madagascar, in the Western Indian Ocean. Four of these islands, i.e., Europa, Bassas da India (an atoll), Juan de Nova, and the Glorieuses Archipelago (MPA since 2012; République Française 2012) are located in the Mozambique Channel, while Tromelin — which is jointly managed with Mauritius (Anon. 2010; Juppé 2012) — is located northeast of Madagascar (Figure 1). Overall, the Exclusive Economic Zone (EEZ; declared in 1978; République Française 1978) of these entities reaches over 640,000 km2 (Anon. 2011), i.e., more than the surface of France's mainland. However, neighbouring countries claim all of these islands: all but Tromelin are claimed by Madagascar, and as well, by Mauritius; the Glorieuses Archipelago is claimed by Comoros and, until 2001, also by the Seychelles (République Française 2001).Tromelin was the first of the islands to be claimed by France in 1776 (Malick 1976), and by the end of the 19th century, all of them were under French rule (Anon. 2011). In 1960, the Îles Éparses became administered by the French island of La Réunion, located east of Madagascar (République Française 1960). In 2007, the Îles Éparses eventually became a district of the Terres Australes et Antarctiques Françaises ('Territory of the French Southern and Antarctic Lands'; République Française 2007), along with the islands of Kerguelen, St Paul & Amsterdam, and Crozet.1The Îles Éparses have virtually always been uninhabited (or occupied for very short periods of time), but they host early-warning meteorological stations since 1950 (this region is under cyclonic threats part of the year), as well as a small but continuous military (and sometimes scientific) presence since 1973 (IUCN 2003; d'Aboville 2007).2 This presence is used to exert sovereignty on these islands, as well as to deter illegal fishing from the coast (Anon. 2011). *  Cite as: Le Manach F and Pauly D (2015) First estimate of unreported catch in the French Îles Éparses, 1950–2010. Pp. 27–35 In Le Manach F and Pauly D (eds.) Fisheries catch reconstructions in the Western Indian Ocean, 1950–2010. Fisheries Centre Research Reports 23(2). Fisheries Centre, University of British Columbia [ISSN 1198–6727].1  Adélie Land on the Antarctic continent also belongs to the same district, with Article IV of the Antarctic Treaty suspending all territorial claims in Antarctica (Guyomard 2010).2  Note that there was a French guano industry occurring in Juan de Nova from the late 19th century until 1972 to supply the Seychellois market (d'Aboville 2007; IUCN 2003). A coconut plantation also exists on the Glorieuses Archipelago; it was planted in the late 19th century and exploited until 1958 by Seychellois mandated by the French government (Malick 1976).Tromelin EuropaJuan deNova 0 500 kmShelfEEZ boundary±GlorieusesBassasda IndiaFigure 1.  Map of the Îles Éparses showing the extent of their EEZs, as well as the -200 m isobaths (i.e., the 'shelf') in the region. 28Patrols are also carried out regularly in the Mozambique Chanel by the French Navy to prevent illegal activities in the French EEZs.Thanks to their remoteness and their uninhabited status, the waters of the Îles Éparses are in an almost pristine state and host a very rich marine diversity (Le Corre and Safford 2001; Perillo 2008), compared, e.g., to its heavily populated and exploited Malagasy neighbour. An extensive mangrove of 700 ha (similar to the one found in Mayotte) is found on Europa (Barnaud 2011; Mangion et al. 2012), which largely motivated its classification as a RAMSAR site in 2011 (Barnaud 2011; RAMSAR 2014). There is also a much smaller mangrove on Juan de Nova, and minor seagrass meadows on two islands. Since 1975, all islands but Juan de Nova have benefited from a status of natural reserve (République Française 1975), which aims to protect the rich flora and fauna, including turtles, cetaceans, coral reefs, and seabirds (Anon. 2011; Quod et al. 2007). Moreover, recent legislation prohibits fishing activities within the 12 nm zone (10 nm around Geyser Bank; 24 nm for purse-seiners; République Française 2010a, 2013a).3Unlike their surrounding waters, the terrestrial parts of these islands have faced a high pressure due to the introduction of various exotic species over time — including rats, cats, goats, chicken, and many plants — which have negatively impacted the indigenous species. Some of these invasive species have been entirely extirpated, while this is still in process for some others (IUCN 2003; Anon. 2011). Given that these islands are uninhabited, there are no fisheries data currently estimated and transmitted to the Food and Agriculture Organization of the United Nations (FAO) and made publicly available via the FishStat fisheries database (FAO 2013). In this report, we apply to the Îles Éparses the reconstruction methods developed around principles in Pauly (1998), described in Zeller et al. (2007) and applied worldwide by the Sea Around Us (see e.g., Zeller and Pauly 2007; Zeller and Harper 2009; Harper and Zeller 2012; Harper et al. 2012). Due to the severe lack of catch data, this work of reconstruction was mostly based on Fermi solutions (von Baeyer 1993; Pauly 2010). We aimed to produce a first estimate of marine fisheries catch in these waters by reviewing the existing literature on the topic and estimating the total extraction of marine fish from 1950 to 2010.material and methodsSmall-scale fisheriesRecreational fisheryVirtually all recreational fisheries occur in Bassas da India's EEZ. This is an example of a highly organized unreported, illegal recreational fishery, with over a dozen South African and Mozambican charter companies offering "extreme adventure holiday [to] fish this mythical fishing destination" (www.bassadaindia.com). Boats filled with tourists (mostly from South Africa) are present in the zone at least half of the year, targeting all sorts of large species ranging from tuna and sharks, to Lutjanidae (snappers) and Coryphaena hippurus (dolphinfish) with spearguns, lines, flies, jigs, and other gears.4 Note that some of these entirely unregulated practices — such as walking directly on the reefs (see vimeo.com/41090694), may result in severe damages to the habitats and the local wildlife, on top of the major impact on some fish stocks. In order to avoid fines by the French authorities, a known trick is to use paired boats: when the patrol arrives, tourists are transferred onto the empty boat, while the catch and the gears are kept on the other boat.5 This way, the authorities cannot charge the charter companies with illegal fishing, as there are no proofs that the catch comes from these waters. However, it has to be noted that the French authorities recently improved their legislation to avoid such practices, by prohibiting the possession of fisheries products onboard boats within the no-fishing zone (10 nm around Geyser Bank, 12 nm elsewhere; République Française 2013b).In order to produce a first estimate of this fishery, we considered that there were 20 boats doing each six trips per year (based on www.bassadaindia.com) for the year 2010, and that this number had increased from zero in 1989 to half of the 2010 level by 2005 (and interpolating in between; i.e., we assumed that this activity slowly expanded in its first years, and expanded more quickly in recent years). We also considered that 500 kg of fish were caught during each trip. We believe this is a conservative estimate, as one recreational fisher reported to have caught at least half a dozen 30–40 kg Thunnus albacares (yellowfin tuna), many Carcharhinus longimanus (oceanic whitetip shark) and C. leucas (Zambezi shark), "a few ignobilis kingfish [Caranx ignobilis]", "a number of black kingfish [C. lugubris]6", "some decent sized snapper", "a number of big wahoo [Acanthocybium solandri], […] the biggest estimated at over 30 kg", and "some other reef dwellers" (Milford 2006).Based on this account, we considered that 50% of the catch was comprised of tunas (80% of yellowfin tuna and 20% of other species) and 20 % of selachimorpha (sharks; 80% of oceanic whitetip sharks and 20% of other species of sharks). The remaining 30% were equally distributed among Lutjanidae, Serranidae, Carangidae, Sphyraenidae, Coryphaenidae, and Istiophoridae.3 However, we know that Geyser Bank has been regularly fished by barques since at least 1997. Thus, unless there really is strict enforcement, these measures may not 'mostly protect' these waters.4  Some of these fish are released, but most are kept for consumption and we assumed 100 % mortality for all species except sharks, for which we assumed 30% survival (based on Diaz and Serafy 2005, Campana et al. 2009, and Butcher et al. 2014).5  The lead author heard this story several times during a trip to South Africa in 2012, while inquiring about a potential fishing trip to Bassas da India.6 Assumed to be the South African common name. Source: www.fishbase.org.Fisheries catch reconstructions for the îles éparses — Le Manach and Pauly 29Holothurian fisherySince 2011, another illegal fishery started to operate in the Îles Éparses' EEZ, from a base in Madagascar (Anon. 2013, 2014a; Pruffer 2013). What started as a small-scale fishery is increasing in organization and size with large (15+ m) mother ships deploying motorized barques and pirogues around  the Glorieuses Archipelago (Geyser and other lagoons) and Juan de Nova (Anon. 2014a).7 These fishers mostly target holothurians while scuba diving, although there is an ancillary catch of sharks (fins and tails kept for the Chinese market as well) and reef fish using lines and spearguns (Anon. 2014a).Fishers likely started fishing these grounds in the early 2000s, i.e., when signs of over-exploitation of Malagasy holothurians started to be conspicuous (Le Manach et al. 2011, 2012, 2013). Noteworthy, it seems that part of this fishery is using boats owned by French expatriates who live in northwest Madagascar part of the year (and get Malagasy people to look after their boats the rest of the year). It happened several times that private sailboats were arrested by the French (or the Seychellois in their own EEZ) with Malagasy fishers and hundreds of holothurians on board, without the owner of the boat knowing that it was no longer moored in Madagascar (Pruffer 2013; G. Cripps, pers. comm. Blue Ventures Conservation).Given that this fishery only started in 2011, i.e., after the end of the time-period studied here, reconstructed catches are not included in the present report. However, for future references, at least ten Malagasy fishing operations (with several boats involved in each) can be assumed to have ventured in the Glorieuses Archipelago and Juan de Nova to catch holothurians in 2013 and early 2014. To get an idea of the catches, one operation that was blocked by the French authorities had collected around one tonne (wet weight) of holothurians (for three small-scale and two semi-industrial boats and over 100 fishers; Anon. 2014a).Reef fisheryFrom late 1989 to mid-1992, a semi-industrial exploitation of reef fish occurred around the Geyser Bank (Glorieuses Archipelago's EEZ) with the 12 m long YVALANN (see Doherty et al., this volume). It quickly stopped due to plummeting catches of the main target species, Lutjanus bohar (two-spot red snapper; Maggiorani et al. 1994; Chabanet et al. 2002). Maggiorani et al. (1994) provided catch data as well as a taxonomic breakdown. Since this vessel used handlines, we considered that all of the bycatch was released in good condition; therefore, we did not estimate any dead discards.Since 1997, fishers from Mayotte also started to travel further offshore to satisfy the local demand for reef fish and they reached the Glorieuses Archipelago's EEZ to target reef fish and some pelagic species (Wendling and Le Calvé 1999; Herfaut 2005; Thomassin and Andrefouet 2009; Fraisse 2010; Doherty et al., this volume). These French fishers operate mostly illegally: fishing activities are forbidden within 10 nm of Geyser Bank and 12 nm of the other emerged land of the Glorieuses Archipelago (République Française 2010a), except for vessels that are deemed safe-enough to travel so far and which can apply for an exemption.8 However, only one vessel was granted this authorization since 2009 (one longliner of 12+ meters), which did not declare any catch (making it illegal with regards to the exemption). As such, all other boats from Mayotte should be considered illegal. Most of this fishery occurs around Geyser Bank, but some barques (which mostly use longlines at night and other types of lines during the day) also travel further and operate around the main islands of the Archipelago.During the first couple of years, it was reported that up to one tonne of fish could be caught per boat and per trip, but this yield quickly declined to only 200–300 kg by the mid-2000s (for longer trips; Thomassin and Andréfouët 2009).9 Therefore, after a strong increase in the number of visits around Geyser bank, a decrease in the number of boats was observed due to this decreasing catch per trip (Quod 2007). Unofficial figures for 2012 suggest that at least 35 barques in Mayotte were equipped for fishing at Geyser and other offshore banks (Doherty et al., this volume).To reconstruct this sector, we considered that the number of barques slowly increased from zero in 1996 to 20 in 2000, and then more rapidly to 60 in 2005. We then considered that this number was halved by 2010, due to the decreasing catch. Regarding the catch, we considered that one tonne was caught by boat and by trip (one trip per month for each barque until 2005, and only 10 per year after 2009) during the first two years, and that this figure declined to 250 kg by boat and by trip after 2005. Regarding the taxonomic composition, we used the same as that of the YVALANN catch published by Maggiorani et al. (1994).Other fisheriesOther very anecdotal small-scale fisheries may occur in the Îles Éparses' EEZ, such as the ones carried by sailboats in transit, military detachments, or even civilians staying at the islands' stations. However, regarding the latter two, it has to be noted that such activities are neither authorized by the hierarchy, nor by the Terres Australes et 7  Such fishing operations have also been reported once in Bassas da India in 2013.8  France is becoming rather worried about such fishery, as the target species (Lutjanidae) are known to be highly sensitive to fishing. The state of the resource is supposed to be assessed as part of the Regional 10th European Development Fund allocated to local French authorities (Mayotte's Conseil Général and Terres Australes et Antarctiques Françaises) in order to implement a "sustainable use of natural resources in Mayotte and the Îles Éparses" (especially in the perimeter of the two Parc Naturel Marins of Mayotte and the Glorieuses Archipelago).9  These illegal fishers will often stay at sea for several days when fishing around offshore banks and risk dangerous sea conditions as well as explosions (Anon. 2014b), in order to remain profitable (Herfaut 2005; Guézel et al. 2009; Fraisse 2010). Some of these illegal fishers operating within the 12 nm are sometimes caught by the French authorities (Anon. 2014c, b). 30Antarctiques Françaises. Although this may occur from time to time, their catch is therefore likely very low. Due to the elusive nature of these fisheries, no estimates were made here.Large pelagics industrial fisheryThe Îles Éparses are located in the second largest tuna fishing ground in the world (FAO 2012), and as such, are attractive to large-scale industrial vessels interested in pursuing this resource. Catches of these fleets were not reconstructed as part of this report. Rather, they were considered to have been reported to the Indian Ocean Tuna Commission (IOTC), and were dealt with as part of the global reconstruction of large pelagics catches (Le Manach et al. in press). However, we present here a brief summary of these fisheries.French fleetFrench purse-seiners (flagged in France mainland, Mayotte, or La Réunion) and longliners (flagged in La Réunion) are active in the Îles Éparses' EEZ (Laurent-Monpetit et al. 2012; www.taaf.fr/Navires-autorises-293). In order to access fishing grounds of the Terres Australes et Antarctiques Françaises, owners of these French vessels must annually apply for a licence, pay a specific fee to contribute to the observation and surveillance program of the Terres Australes et Antarctiques Françaises since 2010 (République Française 2010b, 2013c), and finally, pay fishing rights since 2013 (none until then; République Française 2013d). Since 2008, these vessels must follow the Terres Australes et Antarctiques Françaises regulations (République Française 2008, 2010c, 2013a, 2014),10 and take on fisheries observers (on average 10–15% of trips are monitored).Foreign fleetsSpanish seiners are also authorized to fish in the Îles Éparses' EEZs as part of a bilateral agreement with France. These vessels are either flagged in Spain or in the Seychelles (see Le Manach et al., this volume; www.taaf.fr/Navires-autorises-293).In 1993, there were also licenses delivered to 28 Taiwanese longliners for a trial period of one year (René et al. 1998). It seems that very few industrial vessels have ventured into the Îles Éparses' EEZs over the past couple of decades to fish illegally,11 as they are generally afraid of the increasing French military presence and are thus more inclined to fish in areas further north (known to be more productive and less tightly monitored; e.g., Kenya, Somalia).resultsOverall, catches in the Îles Éparses' EEZs are estimated to have totalled over 2,800 tonnes between 1989 and 2010 (Figure 2A; mostly in the Glorieuses Archipelago with 84% of the total, Bassas da India representing only 16%). The small-scale artisanal barques from Mayotte represented 76.8% of the total (followed by the recreational and semi-industrial handline fisheries, with 14.0% and 6.7%, respectively; Figure 2A).Regarding the taxonomic breakdown, Lutjanus bohar made up 39.9% of the catch, followed by Serranidae, yellowfin 10  Regulations established by the Terres Australes et Antarctiques Françaises are updated every year in accordance with the best available scientific data and national or regional regulations (e.g., IOTC's recommendations and resolutions). 11  Note, however, that this concept of 'illegal' fishing does not apply for the period prior to 1978, as no EEZ existed (République Française 1978).0100200300Total catch (t)Glorieuses ArchipelagoBassas da India01002003001989 1994 1999 2004 2009Total catch (t)YearABLutjanus boharSerranidaeOthersCarangidaeThunnus albacaresScombridaeFigure 2.  Annual reconstructed catch (t), by A) EEZ and B) taxa. See Appendix Table A1 and Appendix Table A2, respectively, for details.Fisheries catch reconstructions for the îles éparses — Le Manach and Pauly 31tuna, other Scombridae and Carangidae, representing 21.5%, 7.0%, 7.3%, and 5.5% of the catch, respectively. The rest of the catch was composed of various species including sharks, other Lutjanidae, Sphyraenidae, and undetermined groundfishes and pelagic fishes (Figure 2B).disCussionIn this report, we provide a first estimate of total marine fisheries catch in the Îles Éparses' EEZs from 1950 to 2010. While the overwhelming majority of the catch is that of the large pelagics industrial fleets (consisting of longliners active since the early 1950s and purse-seiners active since the early 1980s; not included in this report), more recent fisheries are increasingly targeting vulnerable, nearshore species such as holothurians and reef species of fish in an entirely uncontrolled and unmonitored fashion. Signs of over-exploitation are already visible for some of these stocks (e.g., the ones targeted by the barques fishery in the Glorieuses Archipelago and associated banks), with important decreases reported in catch per unit of effort.While our estimates are based on assumptions and are thus perfectible, we do point out the necessity to dedicate increasing efforts towards improving the monitoring and control of these fisheries. This should include the most recent one for holothurians (not reconstructed here), in order to ensure that the exploitation of the marine resources in the biodiversity sanctuary that are the Îles Éparses remains sustainable (if legal), e.g., by restricting fishing activities to areas where stocks are assessed and monitored.aCknoWledgmentsThis report is a contribution of the Sea Around Us, a collaboration between The University of British Columbia and The Pew Charitable Trusts. The authors would also like to acknowledge the support of the Paul G. Allen Foundation, as well as Cédric Marteau and Emmanuel Reuillard (Terres Australes et Antarctiques Françaises) for their valuable input and for sharing their knowledge during meetings held in La Réunion in January 2013.reFerenCesAnon. (2010) Maurice-France : signature d'un accord cadre de cogestion de l'île Tromelin–14 juin.Anon. (2011) Les Iles Eparses–Europa, Bassas da India, Juan de Nova, Glorieuses, Tromelin. Terres Australes et Antarctiques Françaises, Saint Pierre (France). 26 p.Anon. (2013) Pêche illégale dans les TAAF : un voilier dérouté. LINFO.re, edition of November 20, 2013.Anon. (2014a) Opération de lutte contre la pêche illégale aux îles Eparses. Préfecture de La Réunion, Saint-Denis (France).Anon. (2014b) Pêcheurs arraisonnés aux îles Glorieuses : la polémique enfle. ImazPress Reunion, edition of January 9, 2014.Anon. (2014c) 2 navires de pêche épinglés aux Iles Eparses. LINFO.re, edition of January 6, 2014.Barnaud G (2011) Evaluation de trois propositions de sites Ramsar (France) — Analyse des critères d'identification de zones humides d'importance internationale. Rapport SPN 2011–5, Service du Patrimoine Naturel, Muséum National d'Histoire Naturelle, Paris (France). 39 p.Chabanet P, Tessier E, Durville P, Mulocheau T and René F (2002) Peuplement ichtyologique des bancs de Geyser et Zélée (Océan Indien Occidental). Cybium 26(1): 11–26.d'Aboville G (2007) La pêche et l'aquaculture en outre-mer. Avis et Rapports du Conseil Economique et Social, République Française, Conseil Economique et Social, Paris (France). 168 p.Doherty B, Herfaut J, Le Manach F, Harper S and Zeller D (this volume) Reconstructing domestic marine fisheries in Mayotte from 1950–2010.FAO (2012) The state of world fisheries and aquaculture. Food and Agriculture Organization of the United Nations (FAO), Rome (Italy). 209 p.FAO (2013) FishStatJ — Software for fishery statistical time series. V2.1.1. Food and Agriculture Organization of the United Nations (FAO), Rome (Italy).Fraisse N (2010) Analyse des débarquements de poissons à Mayotte par la COPEMAY de 2007 à 2010. COPEMAY, Parc Naturel Marin de Mayotte. 17 p.Guézel R, Quartararo A, Aboutoihi L, Saindou K, Salaün P, Ybrahim B and Arnaud JP (2009) Richesses de Mayotte — Parc naturel marin de Mayotte — Les hommes et l'océan — Mission d'étude pour la création d'un parc naturel marin à Mayotte. Agence des aires marines protégées, Mamoudzou (Mayotte). 72 p.Guyomard A-I (2010) La protection de l'environnement en Antarctique — Droit international et droit comparé. PhD thesis, University of Université de Nantes, France. 806 + 484 p.Harper S and Zeller D, editors (2012) Fisheries catch reconstructions: islands, part II. Fisheries Centre Research Reports 19 (4). Fisheries Centre, University of British Columbia, Vancouver. 143 p.Harper S, Zylich K, Boonzaier L, Le Manach F, Pauly D and Zeller D, editors (2012) Fisheries catch reconstructions: islands, part III. Fisheries Centre Research Reports 20 (5). Fisheries Centre, University of British Columbia, Vancouver. 134 p. 32Herfaut J (2005) Evaluation de la rentabilité économique de l'exploitation des bancs éloignés (Geyzer, Zelée, Castor...) par la pêcherie artisanale mahoraise. Ingénierie de l'environmnement (INENV), Service des Pêches de la Direction de l'agriculture et de la Forêt, Mamoudzou, Mayotte. 22 p.IUCN (2003) Iles Eparses. pp. 107–115 In Gargominy O (ed.) Biodiversité et conservation en Outre-Mer. Planète Nature — Collection du Comité français pour l'UICN — Groupe DOM-TOM. Comité français pour l'UICN, Paris (France).Juppé A (2012) Projet de loi autorisant l'approbation de l'accord-cadre entre le Gouvernement de la République française et le Gouvernement de la République de Maurice sur la cogestion économique, scientifique et environnementale relative à l'île de Tromelin et à ses espaces maritimes environnants. Session ordinaire de 2011–2012–No 299, Sénat, Paris (France). 39 p.Laurent-Monpetit C, Chavance P, Bourjea J and Chassot E (2012) France-Territoires — Rapport national destiné au Comité scientifique de la Commission des Thons de l'Océan Indien, pour sa réunion annuelle 2012. Indian Ocean Tuna Commission. 25 p.Le Corre M and Safford RJ (2001) La Réunion and Îles Éparses. pp. 693–702 In Evans M and Fishpool L (eds.), Important bird areas in Africa and associated islands: priority sites for conservation. BirdLife International, Cambridge (UK).Le Manach F, Andrianaivojaona C, Oleson KLL, Clausen A and Lange G-M (2013) Natural capital accounting and management of the Malagasy fisheries sector — A technical case study for the WAVES Global Partnership in Madagascar. The World Bank, New-York, NY (USA). 46 p.Le Manach F, Chavance P, Cisneros-Montemayor AM, Lindop A, Padilla A, Schiller L, Zeller D and Pauly D (in press) Global catches of large pelagic fishes, with emphasis on the high seas. In Pauly D and Zeller D (eds.), Global atlas of marine fisheries: ecosystem impacts and analysis. Island Press, Washington, DC (USA).Le Manach F, Gough C, Harris A, Humber F, Harper S and Zeller D (2012) Unreported fishing, hungry people and political turmoil: the recipe for a food security crisis in Madagascar? Marine Policy 36: 218–225.Le Manach F, Gough C, Humber F, Harper S and Zeller D (2011) Reconstruction of total marine fisheries catches for Madagascar (1950–2008). pp. 21–37 In Harper S and Zeller D (eds.), Fisheries catch reconstructions: islands, part II. Fisheries Centre Research Reports 19 (4). University of British Columbia, Vancouver (Canada).Le Manach F, Bach P, Boistol L, Robinson J and Pauly D (this volume) Artisanal Fisheries in the world's second largest tuna fishing ground — Reconstruction of the Seychelles' marine fisheries catch, 1950–2010.Maggiorani F, Minet J-P, Roos D and Tessier E (1994) Diagnostic sur l'état de la pêcherie des bancs Geyser et Zélée (Océan Indien) — Analyse des résultats de pêche du navire YVALANN à Mayotte d'octobre 1989 à juillet 1992. Contrat-cade No. 91/1211331/F, Ifremer and Collectivité Territoriale de Mayotte, Le Port (France).Malick M (1976) Notes sur les Iles Eparses de l'Océan Indien. pp. 75–81 In Guézé P (ed.) Biologie marine et exploitation des ressources de l'Océan Indien occidental. Travaux et Documents de l'ORSTOM 47, Paris (France).Mangion P, Lambs L and Fromard F (2012) Nutrient cycling and use by the species Rhizophora mucronata in two contrasted mangrove settings from the French Scattered Islands, Western Indian Ocean In VLIZ Special Publication. Proceedings of the international conference: meeting on mangrove ecology, functioning and management (MMM3), July 2–6, 2012, Galle (Sri Lanka).Milford N (2006) Bassas da India — A spot on the map in the middle of the ocean. SKI-BOAT 22(5): 61–68.Pauly D (1998) Rationale for reconstructing catch time series. EC Fisheries Cooperation Bulletin 11(2): 4–10.Pauly D (2010) Five easy pieces: how fishing impacts marine ecosystems. Island Press, Washington, DC. 193 p.Perillo T (2008) Iles Eparses. Revue d'Ecologie (La Terre et la Vie) 63: 115–116.Pruffer G (2013) Lutte contre la pêche illégale aux îles Eparses. Mayotte 1ère, edition of November 11, 2013.Quod J-P, Barrère A, Chabanet P, Durville P, Nicet J-B and Garnier R (2007) La situation des récifs coralliens des Iles Eparses françaises de l'Océan Indien. Revue d'Ecologie (La Terre et la Vie) 62: 3–16.Quod J-P, editor (2007) Suivi des peuplements ichtyologiques, des peuplements benthiques sessiles et des peuplements d'invertébrés du banc du Geyser. Comparaison des peuplements ichtyologiques entre 1996 et 2006. Direction de l'Agriculture et de la Forêt, PARETO écoconsult, and Agence pour la Recherche et la VAlorisation Marines (ARVAM), Mamoudzou (France). 107 p.RAMSAR (2014) Ramsar sites in order of their addition to the Ramsar list of wetlands of international importance. 40 p.René F, Poisson F and Tessier E (1998) Evolution de la pêcherie palangrière ciblant l'espadon (Xiphias gladius) à partir de La Réunion. pp. 287–312 In Cayré P and Le Gall JY (eds.), Le thon : enjeux et stratégies pour l'océan Indien. ORSTOM, Paris (France).République Française (1960) Décret n°60–555 du 1 avril 1960 relatif à la situation administrative de certaines îles relevant de la souveraineté de la France. Journal Officiel de la République Française: 5343.République Française (1975) Arrêté préféctoral n° 13/DG/IOI du 18 novembre 1975.République Française (1978) Décret n° 78–146 du 3 février 1978 portant création, en application de la loi du 16 juillet 1976, d'une zone économique au large des côtes des îles Tromelin, Glorieuses, Juan-de-Nova, Europa et Bassas-da-India. Journal Officiel de la République Française: 686–686.République Française (2001) Décret n° 2001–456 du 22 mai 2001 portant publication de la Convention entre le Gouvernement de la République française et le Gouvernement de la République des Seychelles relative à la délimitation de la frontière maritime de la zone économique exclusive et du plateau continental de la France et des Seychelles, signée à Victoria le 19 février 2001.République Française (2007) Loi n° 2007–224 du 21 février 2007 portant dispositions statutaires et institutionnelles relatives à l'outre-mer Journal Officiel de la République Française 0045(2): 3220.Fisheries catch reconstructions for the îles éparses — Le Manach and Pauly 33République Française (2008) Arrêté n° 2008–154 du 17 décembre 2008 prescrivant les règles encadrant l'exercice de la pêche aux thons et autres poissons pélagiques autorisée dans les zones économiques exclusives des îles Éparses (Glorieuses, Juan de Nova, Bassas da Indian, Europa, Tromelin). Journal Officiel des Terres Australes et Antarctiques Françaises 40: 18–28.République Française (2010a) Arrêté n° 2010–151 du 9 décembre 2010 portant interdiction de la pêche dans les eaux territoriales des îles Bassas da India, Europa, Juan de Nova, Glorieuses et dans les 10 milles marins autour du banc de Geyser (district des îles Eparses). Journal Officiel des Terres Australes et Antarctiques Françaises 48: 14–15.République Française (2010b) Arrêté n° 2010–104 du 9 octobre 2010 créant une redevance annuelle de surveillance et d'observation de la pêche dans les zones économiques exclusives des îles Glorieuses, Juan de Nova, Bassas da India et Europa. Journal Officiel des Terres Australes et Antarctiques Françaises.République Française (2010c) Arrêté n° 2010–20 du 22 mars 2010 modifiant l'arrêté n° 2008–154 du 17 décembre 2008, prescrivant les règles encadrant l'exercice de la pêche aux thons et autres poissons pélagiques autorisée dans les zones économiques exclusives des îles Éparses (Glorieuses, Juan de Nova, Bassas da India, Europa, Tromelin). Journal Officiel des Terres Australes et Antarctiques Françaises 45: 15–16.République Française (2012) Décret n° 2012–245 du 22 février 2012 portant création du parc naturel marin des Glorieuses Journal Officiel des Terres Australes et Antarctiques Françaises 53: 21–23.République Française (2013a) Arrêté n°2013–14 du 8 mars 2013 prescrivant les règles encadrant l'exercice de la pêche aux thons et autres poissons pélagiques dans les zones économiques exclusives des Iles Éparses (Glorieuses, Juan de Nova, Bassas da India, Europa, Tromelin). Journal Officiel des Terres Australes et Antarctiques Françaises 57: 15–23.République Française (2013b) Arrêté n° 2013–24 du 19 avril 2013 portant interdiction de détention d'animaux marins et de produits de la mer dans les eaux territoriales des îles Bassas da India, Europa, Juan de Nova, Glorieuses et dans les 10 milles marins autour du banc du Geyser (district des îles Éparses). Journal Officiel des Terres Australes et Antarctiques Françaises 58: 6–6.République Française (2013c) Arrêté n° 2013–06 du 30 janvier 2013, créant une redevance annuelle de surveillance et d'observation de la pêche dans les zones économiques exclusives des îles Glorieuses, Juan de Nova, Bassas da India et Europa. Journal Officiel des Terres Australes et Antarctiques Françaises.République Française (2013d) Arrêté n° 2013–08 du 8 février 2013 fixant le montant des droits assis sur les quantités pêchées de thons dans les zones économiques exclusives françaises de Europa, Bassas da India, Juan de Nova, Glorieuses et Tromelin pour la campagne de pêche 2013. Journal Officiel des Terres Australes et Antarctiques Françaises 57: 13–14.République Française (2014) Arrêté n° 2014–51 du 23 avril 2014 prescrivant les règles encadrant l'exercice de la pêche aux thons et autres poissons pélagiques dans les zones économiques exclusives des îles Éparses (Glorieuses, Juan de Nova, Bassas da India, Europa, Tromelin). Journal Officiel des Terres Australes et Antarctiques Françaises 62: 13–21.Thomassin BA and Andrefouet S (2009) Geyser et Zélée : les sommets d'un même volcan sous-marin. Univers Maoré 12: 22–37.von Baeyer HC (1993) The Fermi solution: essays on science. Random House, New York City, NY (USA). 176 p.Wendling B and Le Calvé S (1999) Technologie et utilisation des DCP ancrés dans les eaux de la collectivité territoriale de Mayotte (archipel des Comores, canal du Mozambique, océan Indien). pp. 261–275 In Le Gall J-Y, Cayré P and Taquet M (eds.), Pêche thonière et dispositifs de concentration de poissonsTrois-Ïlets, Caraïbe-Martinique.Zeller D and Harper S, editors (2009) Fisheries catch reconstructions: islands, part I. Fisheries Centre Research Reports 17 (5). Fisheries Centre, University of British Columbia, Vancouver. 108 p.Zeller D and Pauly D, editors (2007) Reconstruction of marine fisheries catches for key countries and regions (1950–2005). Fisheries Centre Research Reports 15 (2). Fisheries Centre, University of British Columbia, Vancouver (Canada). 163 p.Zeller D, Booth S, Davis GE and Pauly D (2007) Re-estimation of small-scale fishery catches for US flag-associated island areas in the western Pacific: the last 50 years. Fishery Bulletin 105(2): 266–277. 34Appendix Table A1.  Total reconstructed catch (t) by EEZ, 1989–2010.Year Bassas da India Glorieuses Archipelago1989 -  15.6 1990  1.8  73.7 1991  3.5  55.4 1992  5.3  45.7 1993  7.1 -1994  8.8 -1995  10.6 -1996  12.3 -1997  14.1  60.0 1998  15.9  120.0 1999  17.6  160.7 2000  19.4  188.6 2001  21.2  228.0 2002  22.9  246.9 2003  24.7  245.1 2004  26.4  222.9 2005  28.2  180.0 2006  33.8  150.9 2007  39.5  123.8 2008  45.1  98.4 2009  50.8  75.0 2010  56.4  75.0 Fisheries catch reconstructions for the îles éparses — Le Manach and Pauly 35Appendix Table A2.  Total reconstructed catch desaggregated by taxa, 1950–2010.YearCarangidaeCarcharhinus longimanusSelachimorphaGroundfishesLutjanidaeLutjanus boharPelagic fishesScombridaeSerranidaeSphyraenidaeThunnus albacares1989 2.1 -- 1.3 - 8.6 - 1.7  1.9 --1990 3.1  0.2  0.1  8.4  0.1  38.5  0.1  5.1  19.0  0.1  0.8 1991 2.7  0.4  0.1  8.8  0.2  25.5  0.2  4.6  14.7  0.2  1.5 1992 3.0  0.6  0.2  11.7  0.3  18.2  0.3  2.4  11.6  0.3  2.3 1993 0.5  0.8  0.2 - 0.5 - 0.5  0.8  0.5  0.5  3.0 1994 0.6  1.1  0.3 - 0.6 - 0.6  0.9  0.6  0.6  3.8 1995 0.7  1.3  0.3 - 0.7 - 0.7  1.1  0.7  0.7  4.5 1996 0.8  1.5  0.4 - 0.8 - 0.8  1.3  0.8  0.8  5.3 1997 4.1  1.7  0.4  9.5  0.9  28.6  0.9  5.5  15.6  0.9  6.0 1998 7.5  1.9  0.5  19.0  1.0  57.3  1.0  9.6  30.3  1.0  6.8 1999 9.8  2.1  0.5  25.5  1.1  76.7  1.1  12.5  40.4  1.1  7.5 2000 11.4  2.3  0.6  29.9  1.2  90.0  1.2  14.5  47.3  1.2  8.3 2001 13.6  2.5  0.6  36.1  1.4  108.8  1.4  17.3  57.0  1.4  9.0 2002 14.8  2.7  0.7  39.1  1.5  117.8  1.5  18.8  61.8  1.5  9.8 2003 14.8  2.9  0.7  38.9  1.6  117.0  1.6  18.8  61.5  1.6  10.5 2004 13.7  3.2  0.8  35.3  1.7  106.3  1.7  17.6  56.1  1.7  11.3 2005 11.5  3.4  0.8  28.5  1.8  85.9  1.8  14.9  45.8  1.8  12.0 2006 10.3  4.0  1.0  23.9  2.2  72.0  2.2  13.6  39.0  2.2  14.4 2007 9.2  4.7  1.2  19.6  2.5  59.1  2.5  12.4  32.8  2.5  16.8 2008 8.2  5.4  1.3  15.6  2.9  47.0  2.9  11.3  26.9  2.9  19.2 2009 7.3  6.0  1.5  11.9  3.2  35.8  3.2  10.4  21.6  3.2  21.6 2010 7.6  6.7  1.7  11.9  3.6  35.8  3.6  11.0  21.9  3.6  24.0 Fisheries catch reconstruction for Kenya — Le Manach et al. 37tentative reConstruCtion oF kenya's marine Fisheries CatCh, 1950–2010*Frédéric Le Manach,1,2† Caroline A. Abunge,3 Timothy R. McClanahan4 and Daniel Pauly51 Sea Around Us, Fisheries Centre, University of British Columbia, 2202 Main Mall, Vancouver V6T 1Z4, Canada2 Institut de Recherche pour le Développement, UMR212 Ecosystèmes Marins Exploités, Avenue Jean Monnet, CS 30171, 34203 Sète cedex, France3 Wildlife Conservation Society, Kibaki Flats 12, Mombasa, Kenya† Current address: BLOOM Association, 77 rue du Faubourg Saint-Denis, 75010 Paris, Francefredericlemanach@bloomassociation.org; cabunge@wcs.org; tmcclanahan@wcs.org; d.pauly@fisheries.ubc.caabstraCtTotal marine fisheries catches were estimated for Kenya for the 1950–2010 time-period using the catch reconstruction approach developed by the Sea Around Us and applied to coastal countries worldwide. This included catches (including dead discards) of the industrial, artisanal, recreational, and subsistence fishing sectors. The total reconstructed catch for domestic sectors for the 1950–2010 time-period reached almost 985,000 tonnes. This figure is 2.8 times the official catch reported to the Food and Agriculture Organization of the United Nations (FAO). Major taxa caught were Lethrinidae (emperors; 9.0%), Scaridae (parrotfishes; 8.8%), Siganus spp. (rabbitfish; 8.6%), Elasmobranchii (sharks and rays; 5.3%), and Carangidae (jacks; 4.7%). The artisanal sector (i.e., small-scale commercial) was the most prominent, with 64% of the total catch. Unreported landings represented 63% of the total catch, whereas dead discards represented close to 2%.introduCtion Kenya is located on the east coast of Africa between Somalia and Tanzania. Its Exclusive Economic Zone (EEZ) extends over 110,000 km² (97th in the World and declared in 1986; Figure 1). The coast is lined with coral reefs covering over 600 km2 (Spalding et al. 2001), except in the central part, where coral growth is prevented by inputs from the Tana River (Ungwana Bay). Mangrove stands are also abundant, especially in the northern half of the coast (UNEP 1998). Despite these rich habitats, marine fisheries are limited due to a narrow shelf, resulting in a small inshore fishing area (Chuenpagdee et al. 2006),1 where essentially all small-scale fisheries occur. Other factors influence small-scale fisheries, such as the northeast and southeast monsoons (from December-March and May-October, respectively), which further restrict fishing activities to inshore waters when the sea is too rough (Obura 2001a). Consequently, marine fisheries have been estimated to represent only 10% of Kenya's total fish catch (FAO 2012); the vast majority of the total fisheries catch comes from the thriving fisheries of Lake Victoria (one of the most important fishing areas on the African continent; FAO 2001; Anon. 2007).Kenyan marine fisheries have always been critical to food security and livelihoods for coastal communities (Devisse 1989), similarly to many developing countries around the world (see also Zeller et al. 2014). For example, Malleret-King (2000) estimated that fisheries provided 80% of the total income to 70% of some coastal communities. Although the number of fishers increased at a rate of 2% per year during the 1980s (McClanahan et al. 2008), there is now some evidence that traditional fishing activities are declining, while other sectors are developing (e.g., sport fishing). One possible explanation is that tourism-related activities play an increasing role in coastal development (Mangi et al. 2007). Thus, some fishers have found alternative livelihoods: or tourists who wish to do sport fishing or visit marine protected areas for their more diverse fauna and protected habitats (Malleret-King 2000; Obura 2001a; Pitcher and Hollingworth 2007).*  Cite as: Le Manach F, Abunge CA, McClanahan TR and Pauly D (2015) Tentative reconstruction of Kenya's marine fisheries catch, 1950–2010. Pp. 37–51 In Le Manach F and Pauly D (eds.) Fisheries catch reconstructions in the Western Indian Ocean, 1950–2010. Fisheries Centre Research Reports 23(2). Fisheries Centre, University of British Columbia [ISSN 1198–6727].1  Defined as the area between the shoreline and either 200 m depth or 50 km distance from shore, whichever comes first.Figure 1.  Map showing the extent of the Kenyan Exclusive Economic Zone (EEZ) and shelf water (to 200 m depth), as well as the location of the major coastal cities of Kwale, Mombasa, Kilifi, Malindi and Lamu (the limits of these districts are also shown), as well as the North Kenya Banks (dotted line).0 100 kmLamuKenyaBanksMalindiKilifiMombasaKwale±!Tana  rive r!!!! 38Declining marine fisheries catches may also be related to declines in fish abundance. Reefs that sustain small-scale fisheries have been under severe pressure for decades in Kenya (see e.g., Khamala 1971; Muthiga and McClanahan 1987; Obura 2001b; Tuda et al. 2008). Their resources have been heavily exploited, and concerns of over-exploitation have been raised since the 1980s (Weber and Durand 1986; UNEP 1989). This has had impacts on both fish biomass and species composition, as evidenced by a long-term decrease in biomass and an increasing proportion of small, herbivorous species (Kaunda-Arara et al. 2003; McClanahan et al. 2008). A Beach Management Unit (BMU) system was introduced in 2006 to reverse these trends by involving communities in fisheries management (Oluoch and Obura 2008). Several gears such as spearguns and beach seines have also been forbidden, and the number of locally-managed marine protected areas (no-take zones, seasonal closures, or gear restrictions) has increased in the last decade. This shift in fisheries management has already had positive results in fish biomass and diversity (Kaunda-Arara and Rose 2004; Abunge 2011), and may lead to increased resilience for local marine ecosystems in light of changing global climate.Official fisheries statistics provided each year since 1950 to the Food and Agriculture Organization of the United Nations (FAO) consist of four distinct taxonomic groups:• Reef fishes (by far the most important group; e.g., Siganidae, Lethrinidae);• Large pelagics (e.g., tunas, billfishes, and sharks);• Shrimps (i.e., 'natantia');• Other invertebrates (e.g., oysters, squids, octopuses).Although it has been claimed in official reports that landing data were reliable (Nzungi et al. 2008), various researchers have criticized the quality of these data, underlining the poor monitoring of fishing activities along the coast, aggravated by low fishers' compliance (Oduor 1984; de Sousa 1987; Obura 2001a). This was clearly evidenced by a small-scale fisheries' reporting system designed in 1984, which determined that almost twice the officially reported amount was actually caught (Carrara and Coppola 1985). Although these new figures should have been processed and released as early as 1985, the absence of any increase in the official catch time-series documents that this was not done. More recently, McClanahan and Kaunda-Arara (1996) and McClanahan et al. (2008) showed that the actual catch per area was as high as 16 t·km-2·year-1 in some areas, starkly contrasting with the number based on official statistic, which oscillated between 2 and 4 t·km-2·year-1 (Kaunda-Arara et al. 2003). However, the situation is thought to have improved over the last decade (Obura 2001a; Muthiga et al. 2008), notably due to the implementation of frame surveys in 2004 (Republic of Kenya 2004–2012). Unfortunately, the monitoring, control and surveillance capacities are still lacking, as many fishers do not report their catch and official catch data still appear to have an unreported component (UNEP 1998; Mangi et al. 2007; Tuda et al. 2008; Maina 2012).In this report, we apply to Kenya the reconstruction methods developed around principles in Pauly (1998), described in Zeller et al. (2007) and applied worldwide by the Sea Around Us (Zeller and Pauly 2007; Zeller and Harper 2009; Harper and Zeller 2012; Harper et al. 2012; Zeller et al. 2014). We aim to improve the overall quality of fisheries statistics by thoroughly reviewing the available literature and re-estimating the total extraction of marine fish since 1950.materials and methodsPreliminary re-allocation of the catchThe nominal catch provided by the Indian Ocean Tuna Commission (IOTC; www.iotc.org/data/datasets) was used to re-allocate the FAO catch of the large pelagics to various sectors. Given that the FAO dataset clearly includes the catch of the longline fleet (targeting swordfish) from 1980 to 1983, we assumed that the catch of this fleet was included in the FAO data for the entire period. When the IOTC catch for a given taxa was higher than the catch of this taxa reported by FAO, we assumed that it was because it was grouped in a more general taxon (due to low catches). We made the same assumption for the sport fishing fleet, thus we also assumed that at least some recreational (i.e., sport) fishing catches were included in the officially reported data. For both these fleets, the re-allocation of the FAO data was done according to Table 1.The remaining catch of 'IOTC species' was re-allocated to the small-scale coastal fleets. However, we point out that except for 'Elasmobranchii', IOTC and FAO data series exhibit considerable and unexplained discrepancies when compared to each other. However, for consistency reasons and due to the rather unrealistic IOTC series (e.g., mostly flat for K. pelamis; plateauing and then steeply dropping for Scomberomorus commerson), we only used the FAO data here. The 'non-IOTC species' catch reported to FAO was also automatically allocated to the either the reef-gleaning sector ('Brachyura', 'Crassostrea spp.', 'Crustacea', 'Holothuroidea', and 50% of 'Octopodidae') or the small-scale coastal fleet (remaining taxa).As a result, the FAO catch was reallocated to several sectors, which were then studied and reconstructed separately (Figure 2).Fisheries catch reconstruction for Kenya — Le Manach et al. 39Table 1.  Correspondence between IOTC taxa and their FAO names, from which their catch was reallocated.Fleet Original IOTC taxon Reallocated FAO taxon PeriodLongline (targeting swordfish) Acanthocybium solandri PerciformesAll years these species were reportedAlopiasElasmobranchiiCarcharhinidaeCarcharhinus falciformisC. longimanusC. obscurusIstiompax indica Makaira indica 1980–83Istiophoridae 2005 onwardIstiophoridae All years these species were reportedIstiophorus platypterus Istiophorus platypterus 1981–83Istiophoridae 2005 onwardIsurus oxyrinchus ElasmobranchiiAll years these species were reportedI. paucusKajikia audax IstiophoridaeKatsuwonus pelamis Katsuwonus pelamisLamna nasus ElasmobranchiiMakaira nigricans IstiophoridaeMarine fishes not identifieda PerciformesPrionace glauca OsteichthyesbPseudocarcharias kamoharai ElasmobranchiiScombridae ScombroideiSelachimorphaElasmobranchiiSphyrna lewiniS. zygaenaSphyrnidaeTetrapturus angustirostris IstiophoridaeThunnus alalunga Thunnus alalunga 1980–83Perciformes 2005 onwardT. albacares Thunnus albacares 1980–83Perciformes 2007 onwardT. obesus Thunnus obesus 1980–83Perciformes 2005 onwardXiphias gladius Xiphias gladius 1980–83Perciformes 2005–08, 2010Osteichthyes 2009Sport fishing Acanthocybium solandri PerciformesAll years these species were reportedAuxis thazard thazardCarcharhinidae ElasmobranchiiCarcharhinus longimanusEuthynnus affinis PerciformesIstiompax indica 1987, 1990–93, 1995, 2008Osteichthyes 1994Istiophoridae 2006–07, 2009Istiophoridae Perciformes 1995Istiophoridae 2000 onwardIstiophorus platypterus Perciformes 1987, 1989–1993, 1995, 2008Osteichthyes 1994Istiophoridae 1996–2006, 2009–10Isurus oxyrinchus Elasmobranchii All years these species were reportedKajikia audax Perciformes 1987, 1989–1993, 1995Osteichthyes 1994Istiophoridae 1996 onwardKatsuwonus pelamis Katsuwonus pelamis All years these species were reportedMakaira nigricans Perciformes 1987, 1990–93Osteichthyes 1994Istiophoridae 1998 onwardMarine fishes not identifieda PerciformesAll years these species were reportedPrionace glauca ElasmobranchiiScombridae PerciformesSelachimorphaElasmobranchiiSphyrna zygaenaSphyrnidaeThunnus albacaresPerciformesT. obesusXiphias gladiusa Given that the IOTC focuses on large pelagics, we changed this taxon to 'pelagic fishes' in our database.b For some reasons, the catch of that species were very high compared to the other species of sharks. Given that the catch of Prionace glauca was even higher than the total catch of sharks reported to FAO, we decided to reallocate it from the higher taxon 'Osteichthyes'. 40Domestic fisheriesSmall-scale, coastal fisheriesSmall-scale fisheries represent the bulk of total Kenyan marine fisheries and essentially involve men.2 Many species of fish are targeted, be they demersal reef species or small pelagic species roaming inshore waters, as well as commercially-important invertebrates such as shrimp, octopus, and lobster (Okechi and Polovina 1994; McClanahan and Mangi 2004; Anon. 2007; Maina and Samoilys 2011; Samoilys et al. 2011a,b). A dozen gears are used on a regular basis to target these different species, ranging from spearguns to beach seine and ring nets, and from traps to boat-operated driftnets (Samoilys et al. 2011a).3 The major fishing grounds are found around Lamu, the mouth of the Tana River, Ungwana Bay/Malindi, as well as the Mombasa area and the North Kenya Banks (see Figure 1; Oduor 1984; Fondo 2004; Maina 2012; Munga et al. 2012). Spearguns, which were introduced in the 1970s (McClanahan et al. 1997), are now commonly used by the poorest fishers because they are cheap (McClanahan et al. 2005), similarly to other less efficient gears (Ochiewo 2002). On the other hand, beach seines (now also illegal) are mostly used because their efficiency is higher than that of any other gears (however, their catch is split into more shares as it requires more men). Beach seines capture a high diversity and size range, overlapping with other gears and, by impacting on the recruitment of a wide range of species, impair the functioning of the ecosystems that are exploited (McClanahan and Mangi 2004; McClanahan et al. 2005; Mangi and Roberts 2006). Due to these different uses, numerous conflicts between gear users have been reported over access to the resource (McClanahan et al. 2005; Mangi et al. 2007; Munga et al. 2010; Fulanda et al. 2011).The pelagic component4 of the small-scale fleet (motorized boats) seems to be increasingly important due to the decline of reef fish, although this fleet is mostly active during the north-east monsoon (when non-motorized boats cannot leave the inshore area; Maina 2012). During this season, fishers that are usually active further offshore are also known to retarget to valuable invertebrate species such as lobsters, holothurians and shells (Marshall et al. 1999; Maina and Samoilys 2011).To re-estimate the total small-scale coastal fisheries, we first estimated the number of fishers from 1950 to 2010. To our knowledge, no reliable time-series of the number of fishers and fishing effort exist for this entire period, although figures have been published by the Government since the early 2000s (Republic of Kenya 2004). Officially, the Government reports that there were approximately 13,000 fishers in 2010 (Republic of Kenya 2012), but Tuda et al. (2008) and Maina (2012) suggested that these numbers were underestimated, and provided a higher figure of 15,000 fishers for as early as the 1990s. To remain conservative, we disregarded these non-official figures, and calculated the ratio of the geometric mean of the number of fishers provided by the Government for the years 2004 to 2010 (Republic of Kenya 2004–2012; the 2010 figure being the average of 2008 and 2012) over the total population in 2007. We then considered this fishers:total population ratio to be constant from 1950 to 2010 and applied it to the total population time-series (Figure 3).5 We further disaggregated this fishers' time-series into five regions (roughly following the 'official districts'; see Figure 1), based on the percentage of the population living in the 15 km-band of each of them.6There also exist some indications that fishers are now active fewer days per year, with an average of 220 fishing days per year (McClanahan and Mangi 2001; Caroline A. Abunge and Timothy R. McClanahan, pers. obs.).7 We therefore assumed that fishers have been active 220 days per year since 1995, but that they used to fish 275 days per year prior to 1975 (i.e., prior to our assumed initial decline in CPUE; see Table 2).2  However, women and children are largely involved in collecting and marketing this fish, and in reef gleaning (see section below).3  Note that dynamite and poison are thought to be rarely used, except near the Tanzanian border (McClanahan et al. 2005).4  Besides medium to large pelagics such as tuna and billfishes, these offshore fishers also target sharks. They are valued as a cheap source of meat (traded up to 100 km inland) and for their dried fins exported to Asia. According to Marshall (1997), Kenya exported at least 140 t of shark fins between 1986 and 1990. Most of these exports (75%) were actually re-exports, as about 10 to 20 t (and 50 t during summer) were imported from Somalia every month. However, anecdotal evidence suggests that the domestic fishery is increasingly widespread, threatening shark populations (Spooner 2012).5  This resulted in an annual growth rate higher than the one published by McClanahan et al. (2008), i.e., +2% per year, but our estimate resulted in a more conservative number of fishers in the earlier period.6  These figures are based on the World Resource Institute's high-resolution GIS files (http://www.wri.org/publication/content/9291), from which we extracted the population living in the 15-km coastal band (assuming homogeneous distribution of the population within each polygon). 7  Note that this number is an average for the entire coastline. Some sources tend to indicate that there are more fishing days in the south, e.g., 300 fishing days (Crona et al. 2010). -         2       4       6       8       10     1950 1960 1970 1980 1990 2000 2010 Catch reported to FAO (thousand t) Year Longline (targeting swordfish) Sport-fishing fleet Small-scale, coastal fleet Reef gleaners Shrimp trawlers Figure 2.  Catch reported to the FAO and reallocated to various fisheries sectors.Fisheries catch reconstruction for Kenya — Le Manach et al. 41We then estimated a CPUE time-series for each of the five regions, based on data collected by the Kenyan branch of the Wildlife Conservation Society (WCS) since 1995, coupled with other anchor points found in the literature (Grottanelli 1955; Samoilys et al. 2011b,c), as described in Table 2.By multiplying these regional CPUE time-series by the number of fishing days and their respective time-series of fishers (Figure 3), we obtained the total 1950–2010 small-scale, coastal fisheries catch (catches in the Tana River estuary were estimated separately; see below).With regards to the taxonomic breakdown, we first adjusted the 1950–1974 data reported to FAO by reallocating part of the meaningless 'Osteichthyes' taxon to the various taxa reported in following years:• From 1972 to 1974, we applied the 1975–79 average taxonomic breakdown of the small-scale coastal fleet minus 'Elasmobranchii' and 'Panulirus sp.' (already reported);• For 1970 and 1971, we applied the updated 1972–76 average taxonomic breakdown of the small-scale coastal fleet minus 'Panulirus sp.' (already reported);• From 1950 to 1969, we applied the updated 1970–74 average taxonomic breakdown of the small-scale coastal fleet;• Finally, we added a new taxon, Scaridae, which seems to represent an important part of the catch according to WCS surveys, but which is absent from FAO data. For this taxon, we simply considered that it was making up 50% of the remaining groundfishes (FAO name is 'demersal perciformes) throughout the time-series.8Once these adjustments were made, we applied the same taxonomic breakdown to the unreported landings estimated above (equals 'total reconstructed' minus 'total reported').The last step was to allocate the total catch to either the subsistence or artisanal (i.e., commercial) sectors. Based on personal observations and communications with local fishers, we considered that 80% of 'Clupeoids' and 50% of 'Carangidae' and the larger groups of 'groundfishes', 'marine fishes not identified', and 'pelagic fishes' were kept for subsistence purposes, while 90% of all other taxa were sold (commercially-valuable taxa corresponding to the 'artisanal catch'; Maina 2012) and the remaining 10% (e.g., juveniles and low-value species) were kept for subsistence.Catches in the Tana River estuary were estimated separately, using shrimp and associated fish catches reported by Munga et al. (2012):8 The rest of the taxonomic breakdown was kept as is for the 1975–2010 period.Figure 3.  Suggested time-series of the total number of fishers (solid line), given the demography of Kenya. The solid dots represent the estimates of the Government (Republic of Kenya 2004–2012; the 2010 point being the average of 2008 and 2012 data). The white square represents Maina (2012)'s estimate, and is provided here as an illustration only. The dashed lines represent the estimated number of fishers in each region (see Figure 1).Table 2.  Summary of the methods used to reconstruct the catch of the small-scale coastal fisheries in the five regions defined in Figure 1, 1950 to 2010.Region Period CPUE (kg·fisher-1·day-1) Note ReferencesMombasa 1950 16.4 Anchor point; assumed 20% higher than anchor point in 1985 Grottanelli (1955)a1951–1974 16.4 Assumed similar to 1950 -1975–1984 16.2→13.9 Linear interpolation until 1985 -1985 13.7 Anchor point Samoilys et al. (2011b,c) 1986–1994 12.8→5.8 Linear interpolation until 1995 -1995–2010 4.9→2.5→3.2 Anchor points WCS dataLamu 1950–1985 16.4→13.7 Similar to Mombasa -1986–2000 13.5→10.0 1975–1985 trend carried forwardb -2001–2010 10.1→11.3 Increase of 1.2% per yearc -Kilifi/Kwale 1950–2000 16.4→2.5 Similar to Mombasa -2001–2010 2.6→2.9 Increase of 1.2% per yearc -Malindi 1950–2010 16.4→7.1 Average between Lamu and Kilifi/Kwale (central position) -a This author does not provide any specifics, but based on his observations, it can be assumed that catches were abundant. Since there were already signs of over-exploitation of Kenyan reefs in the 1980s (Weber and Durand 1986; UNEP 1989), we assumed that the average CPUE was 20% higher than that reported by Samoilys et al. (2011b,c) for the mid-1980s.b We considered that the CPUE decrease in the area of Lamu was slower than in the area of Mombasa, due to a much lower population density, and thus, fishing pressure.c We considered that the trend in CPUE reversed after 2000 as well, similarly to Mombasa area. However, we considered that the recovery rate was half that of Mombasa's, due to lower enforcement of management measures.- 4 8 12 16 1950 1960 1970 1980 1990 2000 2010 Total number of fishers (thousand) Year KwaleMombasaKilifiMalindiLamu 42• From 1963 to 1979, we considered that the entire shrimp catch reported to FAO was small-scale, as the industrial fishery started in 1980. We then applied the average 1963–1979 catch (i.e., 115 t per year) back to 1950, and then from 1980 to 2000;• From 2000 to 2006, we used Munga et al. (2012)'s small-scale catch data;• From 2007 to 2010 (no industrial fishery), we took whichever data was the highest for any given year, between Munga et al. (2012)'s small-scale catch data and FAO data;• Munga et al. (2012) further estimated that fish were making between 87.6% and 93.5% of the small-scale catch from 2001 to 2008. We therefore applied these percentages from 2001 to 2008, and their average (i.e., 90.9%) from 1950 to 2000 and from 2009 to 2010, to estimate the fish catch by small-scale fishers in the Tana River area.We considered that the species of shrimp caught by the small-scale fishers were similar to those targeted by the industrial fleet (see below; Mutagyera 1984), i.e., Penaeus indicus (70.6%), Metapenaeus monoceros (15.6%), P. monodon (5.6%), P. semisulcatus (5.6%), and P. japonicus (2.6%). Similarly, we also used the taxonomic composition of the fish catch reported by Munga et al. (2012), which we applied throughout the time-period. We further considered that 80% of Acanthuridae, Cichlidae, Claridae, Clupeidae, and Protopteridae9 were kept for subsistence purposes (low-value fish), and that 90% of sharks, billfishes, Lethrinidae, Lutjanidae, Mugilidae, cephalopods, Palinuridae, Scaridae, Scombridae, Serranidae, and Siganidae were sold (i.e., artisanal catch; the rest being kept for subsistence). The remaining groups (i.e., Carangidae, and mixed demersals and pelagics) were allocated to the subsistence and artisanal sectors in equal proportions.Industrial shrimp fisheriesThe shrimp fishery is the only sector with a management plan in Kenya (Republic of Kenya 2011; Maina 2012).10 The single shrimp fishing ground of commercial importance is located in the Ungwana Bay (at the mouth of the Tana River; Mwatha 2002),11 and is in fact one of the largest in east Africa (Fulanda et al. 2011). Due to important discharge of sediments and nutrient-rich freshwater from the river, the Ungwana Bay is also known as the most productive fishing ground along the Kenyan coast (Kitheka 2002; Mwangi 2002). A small fleet fluctuating between four and 20 industrial trawlers was active since the late 1970s (Mwatha 2002), but official statistics were only reported since the mid-1980s (Fulanda et al. 2011; Munga et al. 2012). Industrial trawling was restricted to waters beyond nine kilometers from shore, whereas small-scale fishers (who not only target shrimp) were allowed to fish within the 9 km zone.12 However, increasing tensions between the two sectors (e.g., due to gear destruction and resource-sharing (Mwatha 2002; Ochiewo 2002)) forced the government to implement seasonal closures for the industrial fishery in 2001 (Gazette No 7565 of October 31, 2001) and completely ban industrial trawling in 2006.The main targeted species were Penaeus indicus (70.6%), Metapenaeus monoceros (15.6%), P. monodon (5.6%), P. semisulcatus (5.6%), and P. japonicus (2.6%; Mutagyera 1984).To reconstruct the full time-series of industrial shrimp catches, we used the following methodology:• From 1981 to 2000, we subtracted the average 1963–1979 small-scale catch (i.e., 115 t; see section above) from the reported FAO data, in order to estimate the industrial component. The missing 1980 data were replaced by the average between the assumed zero in 1979 (considered to be the last year before industrial trawling for shrimp started) and the 1981 value;• From 2001 to 2006, we took whichever data were the highest for any given year, between Munga et al. (2012)'s industrial data and FAO data.Fish accounted for between 25.6% to 56.7% of the trawlers' total catch from 2001 to 2006 (Munga et al. 2012), and were as high as 70–80% of the total catch before the 2000s (Ochiewo 2002). Therefore, we considered that shrimp were only contributing 20% of the reconstructed total industrial trawler catch from 1980 to 2000, and then used the data provided by Munga et al. (2012) from 2001 to 2006. We also applied the taxonomic composition provided by Munga et al. (2012) from 1980 to 2006.Furthermore, Mwatha (2002) suggested that only adults of commercially-important bycatch species were retained. We assumed 25% of the following species were juveniles and thus discarded: Carangidae, Istiophoridae, Lethrinidae, Lutjanidae, octopodiformes, Palinuridae, Scaridae, Scombridae, Serranidae, and Siganidae. We also assumed that 80% of the 'mixed demersals' were discarded. We applied these ratios from 1980 to 2006, the only exception being Claridae13 (i.e., catfishes), for which we considered that 100% were discarded until 1999, and then only 80% from 2000 to 2006 (Mwatha 2002). The sum of these assumed discard rates applied to the taxonomic breakdown described above amounted to 79.8% of the bycatch being discarded from 1980 to 1999 and 75.1% from 2000 to 2006, in the same range as discards of 67% reported by Kelleher (2005) and the 1:7 shrimp:discard ratio reported by Mwatha (2002).9 This taxon was probably misidentified as it refers to lungfishes, which are strictly limnic.10  There have also been unsuccessful attempts of deep-water shrimp/lobsters fisheries in the Ungwana Bay, but this was not economically feasible (Mutagyera 1984).11  However, note that some Kenyan trawlers are known to have been fishing shrimp illegally in Somali waters (Anon. 2005a).12  This segregation between these two sectors is due to technology: industrial freezers are equipped with funnel-shaped otter trawls and are 25 to 40 m long (storage capacity of 30 to 350 tonnes; engines from 115 to 1,500 horsepower), while small-scale fishers use dug-out canoes and plank wood canoes, thus limiting their activity to a narrow band along the coast.13 This taxon is included in the 'miscellaneous marine fishes' category the Sea Around Us database, as at the time of writing there was no code for this taxonomic group.Fisheries catch reconstruction for Kenya — Le Manach et al. 43Reef gleaningWomen and children have always been involved in collecting invertebrates such as crabs, holothurians and shells all along the coast at low tide (Grottanelli 1955). Overall, though, the catch of reef-gleaners is thought to be smaller than that of the reef fisheries performed by male fishers (Samoilys et al. 2011b).Shellfish account for the bulk of reef-gleaning catches and are mainly collected for the tourism market (Kimani 1995; Marshall et al. 1999), but concerns of over-exploitation have been voiced since the 1970s (Marshall et al. 1999). Holothurians are exclusively targeted for the export Chinese market, and, similarly to shells, it appears that both the average size and the density of holothurians have decreased over-time. They are now mostly targeted by scuba divers in deeper waters, similar to Madagascar (Le Manach et al. 2012) and Tanzania (Bultel et al. this volume). Crabs (mainly Scylla serrata) are consumed locally and are mainly caught in the north, where most mangroves are located (Mutagyera 1984; Kimani 1995; UNEP 1998).We assumed that the number of gleaners was equivalent to 30% of the intermediate number of male fishers (see Figure 3) from 1950 to 1970, and 20% from 2005 onwards (we linearly interpolated the values). This was based on the assumption that reef-gleaning is becoming less important due to the emergence of alternative livelihoods. We then assumed that each gleaner was active 200 days per year and was catching 4 kg·day-1 in 1950. This catch rate was linearly interpolated to 3 kg·day-1 in 2010, based on the aforementioned signs of over-exploitation.Due to the lack of information on this sector, we used the FAO data corresponding to these taxa, and allocated to this sector (Figure 2) to estimate a taxonomic breakdown for our reconstructed catch. For years without data, we carried backward the average percentage of each taxon, and re-scaled the total to 100%. Finally, we created another category, i.e., 'shells', which was deemed to represent the species collected for the tourism market.Longline (targeting swordfish) fleetAs far back as the 1950s, Kenyan waters have been considered to be productive, and Williams (1956) noted the possibility to develop a troll line fishery. As pointed out by de Sousa (1987), FAO data "include the catches from two [domestic] industrial scale tuna longliners which were operated from Mombasa during the early 1980s" (Figure 2). Although IOTC data display the same trend as the FAO data, they are slightly higher. In our database, the difference was thus included as 'unreported landing with respect to data reported by FAO on behalf of Kenya', since we deemed IOTC data to be more accurate.Since 2005, two industrial longliners targeting swordfish have also been registered in Kenya. In 2010, only one vessel remained (then owned by a Spanish company), before it was highjacked by pirates when it ventured into Somali waters (Anon. 2010; IOTC 2012); this vessel was later transferred to the Atlantic Ocean (Nyongesa Wekesa and Ndegwa 2011). Their catch was estimated by a 'liaison officer' to have declined from 730 and 156 t∙year-1. The catch of these longliners were also re-allocated from the FAO series (see above; Table 1).In this report, we did not re-estimate any discards for this sector. This was done separately as part of the Sea Around Us work on harmonizing worldwide catches of large pelagics (Le Manach et al. in press).Sport fishing fleetKenya has been a tourist destination since at least the 1950s (Williams 1970), but mass tourism started in the 1980s (Weaver 1999; Irandu 2004). This sector is now a pillar of the Kenyan economy (Mangi et al. 2007), as there are currently over 1.6 million tourists visiting Kenya every year (Kenya National Bureau of Statistics 2010). Most visitors spend part of their stay visiting places such as the Massai-Mara,Tsavo and Ambosseli National Parks for safaris (Weaver 1999), and about one third also visit coastal areas Williams 1970; Kimani 1995).Kenyan sport fishing started in the 1950s (Williams 1970) and became much more prominent in the mid-1980s (Marshall 1997), due to increased tourism. According to Marshall (1997), there were about 60 sport fishing boats (5–12m long; 60 to 200 trips each per year) that were registered in the late 1990s, but we can expect this figure to have greatly increased in the 2000s. Indeed, Ndegwa (2010) reported that about 30 centers were registered along the coast in the late 2000s; thus, it is easily imaginable that each center has, on average, more than only two boats. As a matter of fact, Ndegwa (2010) also reports that there are on average nine boats per day at sea at Malindi's resort.Sport fishing mostly occurs from April to August, the weather being too rough the rest of the year (Abuodha 1999). Boats mainly use hook and line, in contrast with shore-based recreational fishing (mostly trolling, drifting, and spinning).14 The sport fishing charters generally operate from all major ports and fish the more distant Kenyan Banks, 35–55 km offshore (Ndegwa 2011; Figure 1). However, it seems that, although resorts occur along the entire Kenyan coast, the resorts of Watamu, Malindi (and offshore Kenya banks), Shimoni and Mombasa make up most of the sport fishing activity (Abuodha 1999; Ndegwa 2010). Ndegwa (2010) reports that 22,000 trips were recorded between 1990 and 2008 in the resort of Malindi alone. This author notes, however, a decrease from 1,600 trips per year in the early 1990s to currently 1,200 (Ndegwa 2010).Some authors previously believed that FAO data included some recreational fisheries data at some point in the past (de Sousa 1987), but this was later questioned by Ndegwa (2010). According to the latter author, the Kenyan 14 Although a tag and release project was introduced in 1987 (Abuodha 1999), it seems that most fish are still sold on local markets. When skippers judge the fish to be in good-enough physiological condition, though, they may release it after the photo-shoot.  44Fisheries Department collected sport fisheries data since 1940, but never computerized them. In 2006, the Indian Ocean Tuna Commission and the Overseas Fishery Cooperation Foundation aimed to collect these data in order to create a historical database and analyze CPUE trends. This database is now available at and has been included in the IOTC catch database. Here, we assumed that these data were now included in the FAO data (Figure 2). These reported catches oscillated between 11 and 182 tonnes and averaged 91 tonnes between 1987 and 2010. However, Ndegwa (2010) reported that recreational catches in Malindi's resort alone consistently ranged around 100 t∙year-1, making it therefore likely that only a subset of total recreational catches were ever included in the IOTC dataset.15 As a matter of fact, Maina (2012) reported catches around 206 t∙year-1, with 318 t in 2009. He also noted that much remains to be done to improve the quality of these statistics, reinforcing the feeling that official statistics miss a large part of the recreational sector.To reconstruct this sector, we produced a set of assumptions based on data provided by Williams (1970) for the 1960s:• Sailfish were weighting on average 29.5 kg;• Sailfish were making up 30% of the total catch in weight (the author notes that both sailfish and marlin make up a majority of the catch);• Malindi's area was accounting for half of the recreational catch in Kenya.A catch of zero tonnes was set for 1950, and data were linearly interpolated to 1958, the first year for which Williams (1970) presented data.From 1987 to 2006, we used the data published by the IOTC (Ndegwa 2010), filling the gaps with linear interpolations (1988–89 and 2002–04). For 2007–2010, we used the average of the period 1987–2006; excluding interpolations). Further, we considered that this author only managed to collect half of the actual catch in the area of Malindi (Ndegwa [2010] noted that data still needed to be much improved).To scale these results to the entire Kenyan coastline, we considered that Malindi's resort made 50% of the total catch until 1980, and only 25% from 2000 onward (linearly interpolating in between). This was based on the assumption that other resorts gained a larger portion of the total share due to the tourism expansion in the 1980s.The taxonomic breakdown for this sector was based on Abuodha (1999), although some modifications were made to accommodate the data reported to FAO: Istiophorus sp. (30%) and Scombridae (20%); the rest being equally distributed among Sphyraena spp., Scomberomorus commerson, Makaira spp., Acanthocybium solandri, Elasmobranchii, and other pelagic species.16 The unreported landings were calculated by subtracting the data reported to FAO from the data estimated above (Table 3).Foreign fisheriesDistant-water tuna fleetsHistorically, offshore stocks have remained largely unexploited by local fishers (Anon. 1996), but have long been intensively exploited by distant-water fleets (FAO 2007). Indeed, Kenyan waters are located in the productive Mozambique Channel and are therefore host to highly productive tuna fisheries (Tuda et al. 2008).In recent years, dozens of purse seiners and longliners from the Seychelles, Mayotte, Spain, France and Taiwan have been reported to have fishing licenses in Kenya, with however no conclusive information.17 For example, FAO reported licenses for 33 purse seiners and 30 longliners (FAO 2007), while National reports stated that 19 and 34 licenses were active in 2008 and 2010 respectively (Sigana 2009; Nyongesa Wekesa and Ndegwa 2011); Signa et al. (2008) on the other hand reported a much higher figure of 116 vessels licensed in 2008. Therefore, it seems that many countries have fishing interests in Kenyan waters, but that they may not be legally present (i.e., licensed) every 15  Pitcher and Hemphill (1989) also collected recreational catch data from 1976 to 1987, showing that several hundred yellowfin tuna (i.e., several tonnes) were caught in the resort of Shimoni alone.16  Noteworthy, it seems that shark sightings decreased over the last few decades (Marshall 1997), similarly to most places in the world.17  Note that the EU and the Government of Kenya have been negotiating the signature of a Fisheries Partnership Agreement for several years (Anon. 2005a, 2014).Table 3.  Correspondence between the reported taxa and the assumed FAO taxa, from which their catch was subtracted to calculate the 'unreported landings'.Reported taxon Assumed FAO taxonAcanthocybium solandri Acanthocybium solandriCarcharhinidaeCarcharhinus longimanusIsurus oxyrinchusPrionace glaucaSelachimorphaSphyrna zygaenaSphyrnidaeElasmobranchiiAuxis thazardEuthynnus affinisKatsuwonus pelamisScombridaeThunnus albacaresT. obesusScombridaeIstiophoridaeIstiophorus platypterus Istiophorus spp.Istiompax indicaKajikia audaxMakaira nigricansMakaira spp.Pelagic fishesXiphias gladius Pelagic fishesSphyraena spp.Scomberomorus commerson No reported catch; All 'unreported landing'Fisheries catch reconstruction for Kenya — Le Manach et al. 45year. Indeed, the lack of monitoring and surveillance capacity is thought to be a major incentive for illegal fishing (up to 160 vessels; i.e., only 20% of tuna vessels are licensed) and underreporting (Anon. 2005a, 2007).Catches of this sector are not presented here. Rather, they are estimated as part of the global large pelagic catch reconstruction conducted by the Sea Around Us (Le Manach et al. in press).Small-scale migrant fishersKenyan waters also host migrant fishers from Tanzania. These fishers, coming from the south during the north-east monsoon season, seek calmer waters and often reach the Malindi area, about halfway up the country. These fishers mainly target highly valuable species such as sharks, Carangidae, Lethrinidae and Siganidae (Crona et al. 2010), mainly using lines in the north, and shark nets in the south. Because of the species they target and their usually higher CPUEs (i.e., 2.2 times higher than domestic artisanal fishers, on average), there are often tensions between the two groups. Here, we considered that migrant fishers had a CPUE 2.2 times higher than local fishers in the regions south of Lamu (Figure 1), that their number was equivalent to 10% of that of the local fishers, and that they fished 300 days per year (Crona et al. 2010).resultsDomestic fisheriesThe total catch by Kenyan fishers is estimated to have been almost 985,000 tonnes from 1950 to 2010, which is 2.8 times the amount reported to FAO (Figure 4). It increased from 9,600 tonnes in 1950 to a peak of nearly 27,000 tonnes in 1985. It then decreased to 12,100 tonnes in 2000 and increased again to 15,900 tonnes in 2010.Overall, artisanal, subsistence, industrial, and recreational catches made up 64%, 27%, 5%, and 4%, respectively (Figure 4).We estimated that the small-scale coastal fisheries (including those in the Tana River estuary) caught in excess of 845,000 t between 1950 and 2010 (86% of the total). Catches increased from around 9,2