Longhurst (1965;1969) defined the species assemblages in the Gulf of Guinea by studying the joint occurrence of constituent species in trawl catches and, except for the 20°C isotherm, did not use other characteristics of the habitat where species were found. Domain (1980) used species abundance, depth and bottom type to define species assemblages in the Guinea Bissau-Mauritania area and found very similar assemblages occupying different depth ranges in relation with geographical changes of the thermal structure of the water masses. In this paper we have followed Longhurst's terminology since most scientists in the region are acquainted with it (Annex 1).
Since these assemblages have been determined through bottom trawl sampling, rocky bottoms have not been properly sampled and it is therefore anticipated that assemblages occupying these areas may not have been well characterized. In the Gulf of Guinea rocky bottoms are found in the very littoral fringe (littoral rocky reefs) and on the deeper half of the continental shelf (fossil holocene coral banks).
Fager and Longhurst (1968) have identified a lutjanid community living, according to the authors, on “holocene coral banks” and distinguished from the sparid community. However, this lutjanid assemblage is also referred to in the same paper as living in warm water while holocene coral banks exist below 50 m and down to the edge of the continental slope in subthermocline cool water. The detailed tables given in their paper show that samples characteristic of the lutjanid community were in fact found from 15 to 50 m essentially and mainly at 30–40 m depth, in waters of 18 to 25°C corresponding in the Gulf of Guinea to the discontinuity layer and therefore were caught most of the time in bathymetric strata shallower than the holocene coral banks. As a consequence we have considered in this paper that all inshore rocky reefs are linked with the lutjanid community as well as holocene coral banks that may occur above the 50 m isobath. Below 50 m depth the coral banks have been also distinguished because of their particular relevance to artisanal fisheries and it is assumed that they contain species of the deep sparid community.
Information about species caught, depth and bottom types, oceanographic factors (mainly temperatures) recorded during trawling surveys carried out during the last twenty years, available bottom and navigation charts have been extensively used to prepare the maps of bottom characteristics, assemblages and species distributions. More details about sources of information can be found under individual country headings.
Despite the fact that the two surveys of a “regional” coverage (Guinean Trawling Survey and FIOLENT surveys) have covered at least partially the area under consideration, their results are only available in a summarized form (Robertson, 1977; Williams, 1968 and 1968a; Zupanovic, 1976) and this has prevented us from making full use of the data. The main limitations affecting the available information are as follows:
navigation charts do not give precise information on depths and bottom types for areas not presenting any danger for navigation, i.e., those deeper than 20–30 m;
some of the available bottom charts, especially the one for Cameroon, are not very precise due to problems of estimating the vessel's position;
some of the trawling surveys took place a long time ago, before the tremendous regional increase of abundance in Balistes carolinensis (in 1970–71). However, limited information on species compositions obtained during the 1979–80 trawling survey off Ghana (Koranteng, 1981) shows that at present the other most frequent demersal species are the same as those found during the 1969–70 survey;
trawling surveys seldom covered waters shallower than 10 m and extensive areas of grounds deeper than 10 m where isolated groups of reefs are found between areas of soft level bottom were badly sampled or not at all. As a consequence, the boundaries of the assemblages to be found in such areas were mainly drawn according to bottom characteristics.
In the littoral areas presenting isolated rocks, the distribution of the lutjanid assemblages has been represented in the maps, despite the fact that there may in fact be a mixture of littoral assemblages due to the patchy distribution of bottom types.
Two different methods were used for drawing the geographical distributions of the assemblages and individual species using information on:
assemblages maps were drawn directly and only from bottom characteristics as given by navigation charts and sediment maps, and the position of the 20°C isotherm recorded during the Guinean Trawling Survey. This type of data was used in the preparation of map 1.2.
assemblages maps were drawn using the bottom types indicated by navigation charts or sediment maps together with additional data on type of bottom, depth, depth of the thermocline, temperature and catch composition (number and weight) recorded during past demersal trawling surveys. These maps (all maps but Map 1.2 should therefore be more reliable as based on more convergent information.
Because of the large number of species caught during trawling surveys and the differences between vessels and fishing gears used, it was considered more convenient to determine the boundaries of the species assemblages from the distribution of the most frequently caught species. Only those species caught in excess of 5 kg/h were taken into account and only information on daytime fishing was used in order to eliminate the influence of diel changes in fish distribution and abundance.
Eurybathic species (as identified in Annex 1) were not considered in the analysis as difficult to link to a particular assemblage.
