The PPEAO experimental fishing dataset: Fish from West African estuaries, lagoons and reservoirs

Abstract Background This paper describes a dataset of fish, crustacean and mollusc occurrences extracted from the “Experimental Fishing” section of the IRD's PPEAO information system. PPEAO stands for “Fish communities and artisanal fisheries of West African estuarine, lagoon and freshwater ecosystems”. This database contains information collected using two different methods: experimental fishing and surveys of the artisanal fisheries that exploit these ecosystems. The database is accessible at http://ppeao.ird.fr. New information The current dataset is available on GBIF.org at 10.15468/ra4voa. It comprises the occurrences of 314 fish, crustacean and mollusc taxa collected in experimental sampling surveys of different aquatic ecosystems in West Africa between 1979 and 2013. Different types of fishing gear were used including purse seines, gill nets and fyke nets. The taxa were identified by IRD scientists or by scientific partners well trained in systematics. Most taxa were identified at species level (97% of cases). This dataset is the result of 213 fishing surveys, 5,362 fishing hauls and 31,709 occurrences (28,428 of fish taxa and 3,281 of crustaceans and molluscs). The number of individuals per species and per haul is included and 80% of occurrences are geolocated.


Introduction General context
More than 12 million people work in fisheries in Africa, including more than 2.3 million women. Fish represents, on average, 22% of the protein intake in sub-Saharan Africa; in some countries, this level exceeds 50%. For example, the food security of the Senegalese population depends largely on fish resources: nearly 70% of the animal protein consumed is derived from the sea. Senegal is the second largest fish-producing country in West Africa and now exceeds 500,000 tonnes extracted per year. Estuarine and lagoon fish assemblages are of great importance for fisheries in West Africa, both being directly exploited by local fishermen and constituting nursery areas for many marine species targeted by coastal fisheries.
Knowledge and monitoring of biodiversity is therefore an important issue in West Africa, both in terms of knowledge of aquatic environments and the assessment of aquatic resources for sustainable use. The ecosystem approach to fisheries advocated by FAO particularly requires this type of information. The dataset presented here is of special importance in this context because it combines both old data (the taxonomy of which have been reviewed in the light of current knowledge) and recent data, thus allowing an approach to the temporal evolution of biodiversity. This database is the largest experimental fisheries catch database for West Africa and covers both a large temporal range (more than three decades) and a large spatial range in West Africa (12 ecosystems in seven countries). It comprises fish occurrences mainly from estuarine and lagoon ecosystems (Ebrié lagoon, Sine Saloum estuary including Bamboung MPA, Fatala estuary, Rio Grande de Buba, Gambia estuary), but also from coastal ecosystems (Bijagos archipelago including Urok Island MPA, Banc d'Arguin National Park and Dangara inlet) and also two reservoir lakes (Manantali and Selingue).
After having developed the PPEAO information system, we then felt it was important to provide it with greater visibility and accessibility for the international scientific community. GBIF.org is probably the major global portal for the provision of this type of data, which justifies the choice made here to use it.

roject description
Title: The project PPEAO "Fish communities and artisanal fisheries of West African estuarine, lagoon and freshwater ecosystems" was started in 2006 by the former IRD research unit "Adaptative responses of fish populations to the environmental pressures" (" Réponses Adaptatives des populations de poissons aux pressions de l'environnement", French acronym RAP).
Ecosystems represented in the PPEAO experimental fishing dataset (in chronological order), including the country, the number of surveys, the number of hauls, the sampling period and the geographical coordinates.
Other collectors: Eric Baran (Guinea), Famara Darboe (Gambia), Itaf Deme-Gningue (Senegal), Papa Samba Diouf (Senegal) Study area description: General spatial coverage: West Africa ( Fig. 1) Coordinates: 4°N and 21°N Latitude; 18°W and 2°W Longitude Design description: The sampling surveys completed and embedded in the PPEAO database had different scientific objectives, for example, describing the fish community of an aquatic ecosystem and its space-time variability, searching for particular species, comparing and calibrating biological indicators, testing new sampling devices, demonstrating fishing techniques or carrying out evaluations. In the first case, a sampling protocol was defined, with particular fishing gear and sites that were regularly sampled during each survey.  (Table  1). The taxa are mainly fish and a few crustaceans and molluscs. The PPEAO experimental fishing dataset: Fish from West African estuaries, ...

Sampling description: Methods:
The sampling unit was the fish haul defined by: a site (station), a date (survey), the fishing gear used. After each fishing haul, all the individuals captured (fish, crustaceans, molluscs) were identified, if possible at the species level. All the individuals captured in the same haul and belonging to the same species were counted.
Sampling: Each geographical entity studied was considered as an ecosystem. In each ecosystem, a number of sampling sites was defined. Data collection consisted of several fishing surveys using one or several fishing gears. The most frequently used fishing gear (58% of occurrences) was a purse seine. It was 250 m long, 18 m deep with a 14 mm mesh size. The purse seine was used blindly, without searching for fish in the defined sampling sites. Other types of fishing gear were used: fyke nets (22.6% of occurrences) to sample juvenile fish, gill nets (12%) in Guinea (in addition to purse seine), Mali and the Urok islands MPA, trawl (5.7%) in the deepest area of the Ebrié lagoon and beach seines (1.6%) in the Sine Saloum and the Bijagos.
Quality control: Each sampling survey was supervised by a scientist trained in systematics. Amongst them were IRD senior scientists who contributed to reference papers or books about West African fish species (e.g. Albaret 2003, Albaret 2017, Albaret and Desfossez 1988, Polidoro et al. 2017, Polidoro et al. 2016. Oumar Sadio, the junior scientist who identified fish species for all the recent surveys, was trained in fish identification during three months in the Royal Museum for Central Africa in Tervuren (Belgium). This enabled precise identification of the collected taxa. Most taxa were identified at species level (97% of cases).

Taxonomic coverage
Description: The main aim of samplings for this dataset was to describe the fish communities. Consequently, fish account for the majority of observed taxa (278 out of 314), but some crustaceans, molluscs and cnidarians were also observed (18, 17 and 1 individual, respectively). Amongst the 278 fish taxa, 261 were identified at the species level, representing 98.6% of the occurrences, 14 at the genus level (0.8% of the occurrences) and 3 at the family level (0.6% of the occurrences  Despite the high variability of the sampling effort in the different countries (from 3 surveys in Mauritania to 93 surveys in Senegal), the total number of families identified per country was stable, ranging from 40 to 60, except in Mali where only 14 families were observed ( Table 3). The number of taxa was also low in Mali (50 taxa), whereas it ranged from 71 in Guinea-Bissau (5 surveys) to 141 in Senegal (93 surveys).
Number of fish families and taxa identified per country. As mentioned above, fish assemblages were the main target of experimental fishing surveys. However, other groups (crustaceans, molluscs and cnidaria) were caught accidentally and, in some ecosystems, recorded in the database. Their taxonomy is given according to the World Register of Marine Species (WoRMS Editorial Board 2019). These 36 taxa are listed in Table 4. They account for a total of 3,821 occurrences, identified at different levels: 74% at species level (24 species), 16.7% at genus level (8 genera), 2.5% at family level (1 family), 4.8% at order level (2 orders) and 1.8% at class level (1 class). Crustaceans in the Malacostraca class accounted for half of the total number of observed taxa other than fish (Fig. 2), while molluscs accounted for almost all the other half (10 Gastropoda, 6 Cephalopoda and 1 Bivalvia), except for one Scyphozoa, only identified at the class level.

Additional information Detailed description of the dataset
We now provide a detailed description of the climate, geography, sampling protocol and main characteristics of the assemblages in each monitored ecosystem in chronological order. In the case of the Sine Saloum delta, which was the target for several research projects from the 1990s to the 2000s, all the sampling periods are included in the same section. Conversely, the Bamboung MPA, situated in the Sine Saloum delta, was the subject of a particular monitoring programme and is described separately. Sampling in 1993 in Guinea-Bissau and monitoring of the Urok Islands MPA between 2011 and 2013 are also described separately because they had different objectives.
The largest coastal lagoon in West Africa extends around Abidjan, over approximately 130 km along the coast of Côte d'Ivoire (Fig. 3). It is located between 5.2° and 5.45° North and 3.7° and 4.8°West. It never exceeds 7 km in width for an average depth of 4.8 m and a surface area of 566 km² (Durand et al. 1994). This lagoon is subject to a double influence: freshwater influence from rainfall and river flow, especially from the Comoé River situated at the East end of the lagoon and marine influence via the Vridi canal which provides permanent access to the sea. There are three distinct seasons: a dry season from January to April, a rainy season from May to August and a flood season from September to December (Durand and Chantraine 1982).
In the 1970s, a comprehensive research programme on the environment and the aquatic resources of the Ebrié lagoon was conducted (Durand et al. 1994). This included a study of the fish assemblages of the lagoon conducted by scientists from Orstom and "Centre de Recherches Océanographiques" of Abidjan, from December 1979 to August 1982. The aim of the study was to describe the distribution and functioning of fish assemblages in relation to changes in environmental conditions in space and over time. The results are summarised in a book on environment and aquatic resources of Côte d'Ivoire (Durand et al. 1994) by  and in a more recent article by Ecoutin et al. (2005). Several other publications, based on these data, describe the ecology and the biology of the main lagoon fish species.
Based on preliminary studies, the Ebrié lagoon was divided into six sectors. The Vridi canal separates two dissymetric parts: sectors I and II are located in the east part and sectors IV to VI in the west part. In the centre, sector III is directly influenced by the Atlantic Ocean via the Vridi canal. In the sampling protocol of the fish assemblages, around 80 sampling sites were defined covering all six sectors.  a purse seine from October 1980 to October 1981 and the results were published by Albaret and Ecoutin (1990).
Sampling of Ebrié lagoon recorded 6,537 fish occurrences, corresponding to 60 families and 113 taxa, all identified at the species level (  Table 5. List of the 20 main fish taxa identified in the Ebrié lagoon, with their total abundance (number of individuals) and their proportion (%). These taxa represent 97.5% of the total number of fish caught. During these sampling campaigns, 16 taxa other than fish were recorded, corresponding to 1,007 occurrences: 69% of the individuals were identified at the species level, 24% at the genus level, 0.5% at the family level and 6% at the class level. They were mainly penaeid shrimps Penaeus notialis, jellyfish (Scyphozoa) and crabs belonging to the genus Callinectes.
The Sine Saloum delta is located 100 km south of Dakar and north of the Gambia River and Casamance estuaries between 13.6° and 14.2° North and 15.8° and 16.8° West (Fig.  4). It has three main branches from north to south: the Saloum, the Diomboss and the Bandiala. These branches are shallow (10 to 15 m deep in the centre of the channel, with a maximum depth of 25 m in some places). The branches are interconnected by a dense network of seawater creeks locally named "bolongs", resulting in two groups of islands, partially covered with mangrove. Eight ecological zones have been defined in the Sine Saloum delta (Diouf 1996): Z1 to Z3 in the Bandiala, Z4 in the Diomboss, considered as homogeneous and Z5 to Z8 in the Saloum. The estuary drains a watershed of 29,720 km². The Sudanese type climate is characterised by two seasons: a short, wet and warm season from July to October and an extended dry season, cool from November to March and warm from April to June. Due to the very low gradient of the watershed and the lack of rain, the Sine Saloum is an inverse estuary, i.e. salinity increases upstream, particularly in the Saloum branch. The fish assemblages in this delta have been studied in several research programmes since the early 1990s.

The 1990s
In the early 1990s, a research programme was conducted in the Sine Saloum delta by Orstom and the "Centre de Recherche Océanographique de Dakar Thiaroye" (CRODT), part of the "Institut Sénégalais de Recherches Agricoles" (ISRA), with the collaboration of the "Ecosystèmes Lagunaires" laboratory belonging to Montpellier University, France. This research programme was funded by Orstom. The dataset contains data collected between 1990 and 1997 by the two components of the programme. The "adult and subadult fish assemblages" component was conducted by Jean-Jacques Albaret and the results were published by Diouf (1996) and later by . For the "juvenile fish assemblages" component, conducted by Guy Vidy (Vidy 2000), the dataset specifically comprises data collected using fyke nets.
Two types of fishing gear were used to sample adult and subadult fish assemblages: beach seines and purse seines. Samplings using a "long" beach seine (180 m long, 9 m deep with a 25 mm mesh) were conducted in April, June and September 1990. In December 1990, a few trials were conducted using a "truncated" beach seine and a 100 m long purse seine. From April 1991 to May 1993, a standardised sampling protocol was used with a purse seine 250 m long, 18 m deep with a 14 mm mesh size. Ten sampling sites were defined covering the eight zones of the Sine Saloum delta. On each sampling occasion, each site was sampled twice, one haul in the middle of the channel and the other near the bank, to obtain a representative picture of the diverse habitats. All the sites were sampled at two or three monthly intervals, except the two sites in zone Z8, upstream of the Saloum, which were sampled only twice: once during the wet season (October 1992) and once during the dry season (February 1993). A few "off-protocol" surveys were also done, using the same purse seine, but without following the standardised protocol. Surveys carried out in July, August and September 1994 only covered small "bolongs" fish assemblages. Finally, a last survey was carried out in January 1997 following the standardised protocol.
The study of juvenile fish assemblages began at the end of 1994 and ended at the beginning of 1997 (Vidy 2000). Only the mangrove area was sampled. After a period of adjustment of the sampling technique and gear, the survey lasted two years. The main sampling gear was a fyke net 2.5 m long and 0.4 m in diameter, inspired by the " capéchades" used in the South of France. The net was completed by two wings each 1.5 m long and by a 4 m long "wall" to guide the fish towards the trap. The mesh size was 6 mm for the wall and 3 mm for the second half of the trap. Sampling took place every month at night during the new moon phase, at flood tide to catch the fish entering the mangrove corridors. Two successive strategies were used. In 1995, three sites were sampled on two consecutive nights, with four traps, giving an eight samples set per site and per month. This period was dedicated to the characterisation of seasonal variability of the juvenile fish community. In 1996, three more sites were added and the six sites were sampled on only The PPEAO experimental fishing dataset: Fish from West African estuaries, ... one night using six traps. The aim of this strategy was to obtain details on the spatial organisation of juvenile fish.

Early 2000s
In the early 2000s, the Sine Saloum delta was chosen by the RAP research unit for a comparative study with the nearby Gambia River estuary. The research programme was funded by IRD. The aim of this comparative study of the fish assemblages between the inverse Sine Saloum estuary and the normal Gambia River estuary was to better understand the impact of a major climatic disturbance, hypersalinity. For this reason, more samplings were conducted in the hypersaline upstream area of the Saloum than in the 1990s.
Concerning the sampling of adult and subadult fish assemblages in the main channels, a preliminary survey was conducted in The main fishing gear used for the sampling of juvenile fish assemblages was a fyke net, which was also used in the Gambia River estuary during the same period (Vidy et al. 2004). The gear comprised two parts: the first was a "wall" 8 m long, 3 m high with an 8 mm mesh size. The "wall" was used to guide the fish towards the second part of the trap, the fyke net itself, 3 m long and 0.4 m in diameter, with two wings each 2.5 m long. The mesh size of the fyke net was 6 mm. For each survey, eight traps were deployed at each site for one night, in or near the mangrove, at a maximum depth of 3 m. Fish sampling was carried out in the new moon phase, at flood tide, in order to catch the fish entering the mangrove corridors.

Years 2005-2006
In 2005 and 2006, scientists of the RAP research unit chose to focus on the biological indicators of the state of health of fish populations and assemblages. The aim was to develop a set of indicators that could be analysed jointly to assess the extent of a disturbance in aquatic ecosystems. This work was conducted in the context of the GIBAO project (Groupement Indicateurs Biologiques d'Afrique de l'Ouest). It was funded by IRD, with financial help from the Sub-Regional Fisheries Commission (CSRP) of West Africa, as part of the Regional Partnership for Coastal and Marine conservation (PRCM).
Targeted operations were conducted in the form of purse seine fish samplings. Six sampling sites were defined, one in the uspstream area of the Bandiala (zone 3), the five others in the Saloum (zones 6, 7 and 8), not including the mouth area. During each survey, between 3 and 9 stations were sampled per site. Samplings were made using a purse seine 250 m long, 18 m deep with a 14 mm mesh size. All the hauls were classified as "off protocol" because they did not cover the entire Sine Saloum delta.

General description of catch data
The consecutive research programmes in the Sine Saloum delta recorded 8,567 fish occurrences, in all the periods combined, corresponding to 56 families and 126 taxa, 95.5% of them being identified at the species level, 2.5% at the genus level and 2% at the family level (  Table 6. List of the 22 main fish taxa identified in the Sine Saloum delta, with their total abundance (number of individuals) and their proportion (%). These taxa represent 98% of the total number of fish caught. During the samplings in the Sine Saloum delta, 969 occurrences representing 14 taxa other than fish were also recorded, of which 66.7% were identified at the species level, 18.3% at the genus level, 7.9% at the family level and 7% at the order level. Most of the occurrences corresponded to shrimps Penaeus notialis and crabs Callinectes amnicola. Some cephalopoda (Loliginidae and Sepia) were also quite often collected.
In order to assess the influence of the riverine system on the estuarine ichthyfauna, a comparative study was conducted in Guinea between January 1993 and March 1994. The aim of the study was to compare the fish assemblages in the estuary of the Fatala River with those in the neighbouring Dangara inlet, located 20 km away and not connected to a river. This study was conducted by Eric Baran for his PhD thesis (Baran 1995), which was directed by Christian Lévêque and advised by Jean-Jacques Albaret. The whole research programme was funded by Orstom, in cooperation with the inland fisheries department of the "Centre National des Sciences Halieutiques de Boussoura" (CNSHB). The study area was located in the coastal zone of Boffa, which belongs to the "Southern Rivers" region ( Fig. 5). It is located between 10° and 10.3° North and 14° and 14.25° West. The tropical wet climate is characterised by a dry season from November to May and a wet season from June to October with peak rainfall in July. Rainfall is very high (approximately 3,500 mm/ year) with the hydrological cycle reaching a maximum in August or September (Baran 1995).
In the Fatala estuary, the estuarine zone is limited by a rock sill 60 km upstream which stops both the dynamic and saline actions of the tide. As the area 10 km upstream of the estuarine zone could not be accessed by boat, sampling was limited to the downstream 50 kilometres. Two fishing gears, a purse seine and gillnets, were used alternately every month, to obtain a comprehensive description of fish assemblage variability. The purse seine (250 m long, 18 m deep with a 14 mm mesh size) was used every two months from January 1993 to January 1994, at eight equidistant sites, one of which was located in the ocean in front of the river mouth. Each site was sampled during daylight, each time with two hauls, one haul in the mid-channel and the other as near as possible to the bank (minimum water level: 2 m). The gillnets were set up in parallel and close to the bank, in order to provide a better sampling of riparian fish assemblages. A gillnet set was composed of ten panels of different mesh size (10, 12.5, 15, 17.5, 20, 22.5, 25, 30, 35 and 40 mm).
Each panel was 25 m long and 2 m deep. Four sampling sites, the same as those used for the purse seine samplings, were located 3, 17, 33 and 46 km from the mouth, respectively. At each site, four "stations" were defined, corresponding to two subsequent nights on both sides of the estuary. In the Dangara inlet, three sites were defined. Dangara surveys took place immediately before or after those in the Fatala estuary, i.e. every two months. Only purse seine samplings were performed, twice at each site as in the Fatala estuary (midchannel and close to the bank).
Together, the samplings performed during this study in Guinea recorded 3,910 fish occurrences, corresponding to 40 families and 86 taxa, 99.8% of which being identified at the species level and 0.2% at the genus level (   Only fish were collected during this study and no information is available about taxa other than fish.
The Bijagos Archipelago is located off the coast of Guinea-Bissau between 10.9° and 11.6°N orth and 15.6° and 16.5° West. The Rio Grande de Buba is a small coastal river flowing Table 7.
List of the 20 main fish taxa identified in the Fatala Estuary and the Dangara inlet, with their total abundance (number of individuals) and their proportion (%). These taxa represent 96% of the total number of fish caught.
in the west central region of Guinea-Bissau, between 11.4° and 11.7° North and 15° and 15.5° West. It flows into the Atlantic Ocean in front of the Bijagos Archipelago (Fig. 6).

Following a request from the "Centre Canadien d'Etudes et de Coopération Internationale"
(CECI), a team from Orstom and the "Centre de Recherche Océanographique de Dakar Thiaroye" (CRODT -Sénégal) carried out a scientific survey in the Bijagos Archipelago at the end of March 1993. This team was composed of Itaf Deme-Gningue and Pape Samba Diouf from CRODT and Jean-Jacques Albaret from Orstom. Scientists and technicians from the Fisheries Department of Guinea-Bissau and from CECI also took part in the survey. Following this survey, a second survey was carried out by the same team in the Rio Grande de Buba at the begining of April 1993 in response to a request from the International Union for Conservation of Nature (IUCN). Jean-Louis Kromer (UICN) participated in this survey. The results of the surveys were published in two scientific reports (Deme-Gningue et al. 1994 andDiouf et al. 1994).
In the Bijagos Archipelago and the Rio Grande de Buba, a sampling plan of 36 and 24 sites, respectively, was followed, considering the diversity of habitats and bathymetry. The GPS positions were recorded for the Bijagos Archipelago, but not for the Rio Grande de Buba. The main fishing gear used was a purse seine (250 m long, 18 m deep with a 14 mm mesh size). At some sites, a beach seine was used to provide information about riparian fish communities. It was used either instead of the purse seine (at sites 25 and 30 in the Bijagos Archipelago) or in addition to it (at sites 32 and 34 in the Bijagos Archipelago and sites 1 and 10 in the Rio Grande de Buba). This beach seine was 50 m long, 2 m depth with a 30 mm size for the wings and 14 mm mesh for the bunt.
The two surveys performed in 1993 in the Bijagos Archipelago and the Rio Grande de Buba in Guinea-Bissau recorded 258 fish occurrences, corresponding to 43 families and 60 taxa; all fish were identified at the species level ( During these two surveys, 28 occurrences of taxa other than fish were identified. They belonged to five taxa, mainly shrimp Penaeus notialis and crabs belonging to the genus Callinectes. Table 8. List of the 20 main fish taxa identified in the Bijagos Archipelago and the Rio Grande de Buba during the 1993 survey, with their total abundance (number of individuals) and their proportion (%). These taxa represent 97.5% of the total number of fish caught.
At the begining of the 2000s, the Gambia River estuary was chosen by the RAP research unit of IRD as a reference ecosystem for a comparison of West African estuarine fish assemblages. The Gambia River estuary has rarely been disturbed by drought and functions like a "normal" estuary, with no upstream dams; it is surrounded by an abundant mangrove in good condition, with little pollution and subject to low fishing pressure.
A joint IRD/Gambian Fisheries Department research programme was conducted from 2001 to 2003, for the "Evaluation of the fish resources and estuarine resources stewardship of the Gambia River". It was funded by the French Department for Cultural Action and Cooperation (SCAC). The study area comprised the lower, middle and upper estuary, from Banjul at the mouth to Deer Island 220 km upstream (Fig. 7). The Gambia estuary is located between 13.2° and 13.7° North and 15° and 16.6° West. In this region, the wet season extends from June to October, with peak rainfall in August. The cool dry season lasts from November to March and the warm dry season from April to June.
Five surveys were conducted to characterise the adult and subadult fish assemblages in the main stream of the Gambia River estuary, covering the whole hydrological cycle. Not all the subsequently sampled sites were visited during the preliminary survey (November 2000), whereas six of the originally envisaged sites were abandoned after the preliminary survey. During each of the four following surveys (June 2001, September 2001, December 2001and April 2002, 44 sites were sampled, only using a purse seine (250 m long, 18 m deep with a 14 mm mesh size). The results were published by  and . protocol to be adjusted. Then, in 2002 and2003, three surveys per year were conducted between May and November, the season of maximum juvenile recruitment, at six sites along the estuary. The main gear used to sample juvenile fish was the same as that used in the Sine Saloum delta during the same period (see above). The results were published by Vidy et al. (2004).
Together, the samplings performed in the Gambia River estuary recorded 3,710 fish occurrences, corresponding to 40 families and 92 taxa, almost all identified at the species level (  Table 9. List of the 20 main fish taxa identified in the Gambia estuary, with their total abundance (number of individuals) and their proportion (%). These taxa represent 95% of the total number of fish caught. During the survey of the Gambia River estuary, 893 occurrences of taxa other than fish were recorded. They corresponded to 10 taxa, mostly Callinectes amnicola and Penaeus notialis.  . Manantali is located between 12.9° and 13.3° North and 10.2° and 10.4° West and Sélingué is located between 11.2° and 11.6° North and 8° and 8.4° West. Three methods of assessment were combined for this study: fish gillnet sampling, surveys of artisanal fisheries and hydroacoustics (Coll et al. 2007). Regarding fish gillnet sampling, three surveys were conducted in both reservoirs: a preliminary survey in June 2002, which enabled adjustment of the sampling protocol and two surveys covering the key hydrological seasons: the end of the dry season (June 2003) and the end of the rainy season (October 2003), corresponding to the minimum and the maximum water levels, respectively. A detailed description of the sampling protocol is given in Laë et al. (2006) and Coll et al. (2007). Three zones were defined a priori in Manantali reservoir: the dam, the middle zone and the upper zone. The same zones were defined in Sélingué reservoir, separating Bale and Sankarani, the two rivers flowing into the lake. Due to limited accessibility, time and high costs, only two areas in each reservoir were actually sampled: the dam and central zones in Manantali and the dam and central zones on the Sankarani side in Sélingué (Coll et al. 2007).

Selingué and Manantali reservoirs (Mali)
Sampling was carried out using monofilament gillnet panels with five different mesh sizes (10, 15, 22.5, 45 and 80 mm) joined together to form a gang. Each single mesh panel was 25 m long and 3 m deep. Each gang therefore comprised a fishing area of 375 m² (Coll et al. 2007).
Together, the samplings performed in the Manantali and Sélingué reservoirs recorded 1,266 fish occurrences, corresponding to 14 families and 50 taxa, almost all identified at the species level (  During these samplings, only fish were collected and no taxa other than fish were recorded.
This dataset is the result of the biological survey of the Bamboung Marine Protected Area (MPA) in Senegal located 130 km south of Dakar, between 13.75° and 13.84° North and 16.5° and 16.56° West, in the Biosphere Reserve of the Sine Saloum Delta (Fig. 4). The Bamboung bolong is a small affluent of the Diomboss, one of the main branches of the Sine Saloum Delta. The MPA covers 68 km², with a wetland zone composed of mangrove and savannah and a water body of approximately 4 km². It is one of the sites chosen for the purpose of demonstration in the "Narou Heuleuk" (tomorrow's share) project. It was proposed and driven by the Oceanium, a Senegalese organization for the protection of marine resources, and originally funded by the French Fund for the World Environment   Table 11.
List of the 20 main fish taxa identified in the Bamboung MPA and its surroundings, with their total abundance (number of individuals) and their proportion (%). These taxa represent 98.6% of the total number of fish caught.
The Banc d'Arguin National Park was sampled as part of the AMPHORE project "Marine Protected Areas: biodiversity conservation tools and sustainable management of fishery resources" ("Les Aires Marines Protégées: outil de conservation de la biodiversité et de gestion durable des ressources halieutiques") funded from 2008 to 2011 by the French National Research Agency (http://www.agence-nationale-recherche.fr/?Projet=ANR-07-BDIV-0009). A purse seine (250 m long, 18 m deep with a 14 mm mesh size) was used to sample fish assemblages by a team of IRD scientists, in cooperation with the Mauritanian Institute of Oceanographic Research and Fisheries (IMROP). Thirty four sites were defined inside and outside the MPA. Some had to be cancelled because they were not accessible by boat. Two surveys were conducted in 2008, in the dry season (7-14 May) and the wet season (8-13 October). A third survey began on 27 April 2010, but had to stop after the first four hauls due to technical problems, before starting again from 22 -28 May 2010. The wet season survey could not be carried out in 2010 (Sadio 2015).
During these samplings, 44 occurrences of 12 taxa other than fish were recorded. They were either identified at the species level (66%) or the genus level (34%). The most frequent taxa belonged to the genera Murex and Sepia and to the species Penaeus notialis.
The Urok Islands complex is located in the coastal area of Guinea-Bissau, in the northern part of the Bijagos Archipelago biosphere reserve (Fig. 10). The Marine Protected Area of the Urok Islands was created in 2005, after agreement was reached between the resident populations, the NGO Tiniguena and the administrative authorities of Guinea-Bissau. List of the 20 main fish taxa identified in the Banc d'Arguin National Park, with their total abundance (number of individuals) and their proportion (%). These taxa represent 98.4% of the total number of fish caught.
The Urok Islands MPA covers 545 km², 398 km² of which is in the maritime zone. It is located between 11.4° and 11.6° North and 15.9° and 16.1° West. The MPA is broken down into three zones, where fishing is authorised at different levels. The aim of zoning was to allow the central zone, where the fishing restrictions are the strictest, to have both attraction and spillover effects. This study was part of the CEPIA project of the Sub-Regional Fisheries Commission (Commission Sous-Régionale des Pêches -CSRP) entitled "Building together a fisheries management integrating MPA" ("Construire Ensemble une gestion des Pêches Integrant les AMP"), funded by the French Development Agency (AFD) and coordinated by the International Union for Conservation of Nature (IUCN). The sampling protocol was designed and implemented by a team of IRD scientists (Sadio 2015). The surveys were conducted in cooperation with the Instituto Nacional de Pesquisa Agraria (INPA) and the NGO Tiniguena, using both bottom and surface gillnets. Each set of gillnets was made up of six panels with different mesh sizes (14, 25, 36, 50, 60 and 80 mm). Each panel was 25 m long and 2 m deep. The gillnets were set up at high tide, with a fishing depth of at least 7 m. The surveys were carried out in November 2011 and 2012 at the end of the rainy season and in May 2013, at the end of the dry season (Sadio 2015).
The three surveys performed in the Urok Islands MPA recorded 403 fish occurrences, corresponding to 30 families and 47 taxa, all identified at the species level (Table 13). Fish assemblages were dominated by the Ariidae family, with Carlarius latiscutatus accounting for 29.9% and Carlarius parkii for 9.9%. The Pristigasteridae Ilisha africana and the Haemulidae Pomadasys perotaei each represented 5.4% of the occurrences. The Carcharhinidae Rhizoprionodon acutus was also often observed (4.6% of the occurrences).  During these samplings, only seven occurrences of taxa other than fish were recorded.