The available data by month and region (or by sector or sampling site), were presented by the experts responsible for the preparatory mission. The report of this mission (Annex 3), enabled the working group to take stock of the data which for each species could be utilized for detailed analysis of the CPUE by the grid method presented above (3.3).
For Sardinella aurita and S. maderensis, it was possible to follow the catches and efforts on a monthly basis and by sector since 1982 for the industrial fishery in Côte d'Ivoire, the artisanal fishery in Ghana and since 1985 for the artisanal fishery in Togo. Similar data (but for sampling site) existed for the artisanal fisheries in Côte d'Ivoire since 1982 at Vridi, then 1987 for the total coastline and since 1987 for Ghana. The semi-industrial and industrial fisheries of Ghana appeared too difficult to follow on a spatia-temporal basis; for a number of strata there were no data because of the frequent change of fishing gear by these boats, especially the semi-industrial ones.
Catches and yields of the anchovy could be well followed from the artisanal fleets of Ghana (since 1982), Benin (since 1985) and Togo (since 1988).
The mackerel appeared less accessible to the artisanal fishery and their analysis was managed from the industrial fishery of Ghana (since 1985) and Côte d'Ivoire (1982).
It should be noted that, because of the limited time available for compilation, the CPUE data by sampling site in Ghana could only be carried out for catches by gear and month (catches by canoes using a given gear on the fishing ground). The significance of such an index should be discussed. The CPUE by region were catches by trip and were directly comparable with the data from the other countries.
The data on length frequencies practically concern only the two sardinellas: they were only complete on a fine spatia-temporal basis for the Ivorian industrial fishery (since 1984); for the other fisheries the very dispersed samples existed in time and in space from 1981 in Ghana and 1987–88 in the other countries.
The following observations are drawn from the qualitative analysis of the grids of yields and catches by sector and by month for 1982–90. These analyses could advantageously be completed using a multivariate statistical approach. The qualitative synthesis of the information permits the effect of fishing gears and the effect of year to be examined here. Figures 9a–9d present certain aspects of the grids used for these analyses.
Effect of gear
Ghana beach-seine: The large catches take by this gear were essentially in the Volta Region; in the eastern part of Ghana, for S. aurita, S. maderensis and Engraulis encrasicolus. The peaks of catches of round sardinella and anchovy were synchronised and spread principally from August to October at the end of the major cold season and, secondly for 1 or 2 months between December and March. Since 1987, a de-seasonalization of catches of S. aurita was observed; this was not the case with anchovies.
On the contrary, S. maderensis was mostly caught in November-January, in the region where the minor upwelling was weak, and then in March-April. However, a delocalization of catches for this species is observed since 1985 with large catches in the western sector, and then in 1987 an extension throughout the coastal region associated with the seasonalization.
Beach-seines and canoes of Benin and Togo: Between 1986 and 1989, the period of abundance of the beach seine for round sardinella and anchovy was relatively synchronized with that in the Volta Region of Ghana. The canoes of the two countries caught few S. aurita. The exploited modal sizes of this species were 9–13 cm (Togo), but individuals of 20–24 cm are also found. For anchovy which is a target species in Togo, the extension in time of periods of high abundance between Benin, Togo and Volta Region in Ghana were not always the same if thus could be explained by a displacement of the epicentre of the stock. There were no catches of S. maderensis by the beach-seine and purse-seine in Benin and Togo. Only the sardinella net caught some large individuals of this species in Benin in sizable quantities, somewhat outside the upwelling season (October-April), and for a short period from May to July.
Poli/ali/watsa nets of Ghana: In contrast to the beach-seine, catches by these nets practically concern the three regions in the west of Ghana. Large catches of S. maderensis were recorded from May to September slightly ahead of the cold season, and of S. aurita and anchovies from June to October. Certain years, in particular since 1985, showed peaks of activities for the three species in the minor cold season (November-February). For the three species, there exist only one or two regions in a given season with high catches and a simultaneous balance is observed for the three species of importance in the regions: from 1982 to 1985 S. aurita and S. maderensis were principally exploited in the Greater Accra Region, when the anchovies were caught in the west in the Central Region; from 1985 to 1987, the catches of S. aurita and S. maderensis were made in the western part of Ghana when not much anchovy was caught; in 1988 and 1989, S. aurita and S. maderensis were caught in the Central Region when the anchovies were caught mostly in the Greater Accra Region (an inverse situation of the years 1982–84).
These observations confirm the evidence of the Ghanaian researchers that anchovies are targeted only when sardinellas are less abundant. The other explanation would be the search for different biotopes between the sardinellas and the anchovy.
The length frequency data available in Ghana showed clearly the effect of mesh selectivity of the ali net on S. aurita and S. maderensis, catching mainly individuals with a mode of 18–20 cm at the end of the upwelling season. This selectivity effect, although less marked, also exists for the purse-seiners which seek mainly large fish (11–20 cm long). Only the poli nets exploit “with eyes closed” all sizes with the principal mode spread over the whole 6–20 cm range.
Ivorian industrial seiners: Strong indices of abundance were obtained from 1982 to 1989 on S. aurita between July and October/ November for sectors 4, 6 and 7, and to a lesser degree in sector 5. For this species large yields were observed in December to March-April in the whole of Côte d'Ivoire for the first time in 1985, then in 1986 and 1987, principally in the two sectors in the west. In 1988 and 1989, a situation similar to that of 1982–84 was again observed. For S. maderensis, two distinct cycles seem to characterize this fishery: in the western sectors (6 and 7) where the catches were most abundant, there appeared to be high yields from August-October to March-April, even though in the east (sector 5) the very wide periods of abundance concerned the months May-July, as in Ghana at the beginning of the major cold season. Also, as in Ghana, the purse-seiners exploited large fish of 17–23 cm length.
Ivorian canoes: The catches concern mostly the regions east (Vridi) and west (Sassandra, Bereby). The periods of great abundance of S. aurita were synchronized with those of the purse-seiners, but their extension was often of unequal duration; sometimes inferior to those of the seiners (1987, west Côte d'Ivoire) or superior (1988–89 west Côte d'Ivoire). For S. maderensis, periods of great abundance were noted more frequently than for S. aurita. Moreover, these periods of great abundance are complementary to those of the seiners, probably implying a phenomenon of accessibility to the resource due to inshore-offshore migration. In the years 1982–84, at Vridi, the species seemed available mostly to canoes, even though it became equally available to the two gears in 1988 and 1989. These gears selectively exploit individuals, the majority of which are between 10 and 20 cm.
Ivorian beach-seines: The data on this gear, essentially in sectors 5 and 7, were poor. The length frequency of S. aurita collected at Sassandra showed modes at 6–9 cm. Seining at Jacqueville showed catches of small-sized S. maderensis from September to March-April (like the artisanal seiners at Vridi).
The effect of year
Three groups of years could be identified in the series studied, and the major changes affecting the different fisheries are described here. The years 1985 and 1987 appear to be the turning point.
1982–84: For the three species, the seasonality of the catches was very much marked both for the beach-seine of Ghana (August-November for S. aurita and Engraulis encrasicolus, November-February for S. maderensis), and the artisanal and industrial seiners of Ghana and Côte d'Ivoire. S. aurita attains a maximum mode of 16–18 cm in Ghana, and 16–19 cm in Côte d'Ivoire. Similarly, the localization of large catches by gear was stable from year to year.
1985–87: there were drastic changes in the fisheries and the following phenomena were observed:
The appearance of large catches or indices of abundance between November and April for all the gears in the region and S. aurita, and for the beach-seine of Ghana with respect to the three species (the two sardinellas and E. encrasicolus). These large abundances correspond to an intensification of the upwelling index of the minor cold season along the entire coastline. This bi-seasonalization associated with the high abundance explains the increased catches during this period. On the contrary, the anchovy disappeared from the landings of canoes in 1985 and 1986, which could be the result of not targeting the species. For the round sardinella, no spectacular changes in length were observed in 1985, even though in 1986 and 1987 the species attained a modal size of 21 cm in Ghana and 23–24 cm in Côte d'Ivoire.
Displacement of the sectors of large abundance: There was a concentration of S. aurita, in Côte d'Ivoire and in western Ghana, and also in the two regions east of Ghana; the Central Region had apparently become a desert for this species. For S. maderensis, these years saw an extension of the sectors of large abundance: sector 6 in Côte d'Ivoire together with all the sectors in Ghana. It was difficult to know if the absence of anchovy in the catch in the two central zones of Ghana was due to an absence of the resource, a migration towards the east (where an increase in abundance to the beach-seine and a de-seasonalization are observed) and Togo or to a problem of species targeting.
1988–1989: This period follows 1987, when the upwelling of the major cold season was exceptionally weak along the entire coastline, and was marked by a clear relaxation of the cooling of the minor cold season.
The indices of abundance of the Ivorian seiners and the Ghanaian canoes are as follows: for S. aurita, and as in 1982–83, a period of great abundance during the major cold season was observed in the whole region in 1988 and essentially in the east of Côte d'Ivoire and the sectors east of Ghana in 1989 (a situation of 1984 and years before 1982). The round sardinellas had length modes at 18–20 cm in Ghana and 19–22 cm in Côte d'Ivoire; for S. maderensis, a reduction in the zones and periods of great abundance were observed, except in the western part of Côte d'Ivoire. For the anchovy, the return of catches in the Central and Greater Accra Regions occurred in 1987 and until 1989 (with an inverse of the situation of the 1982–84 period, a domination of the catches in the Central Region).
For S. aurita, unlike in 1982–84 when the catches of the minor cold season were insignificant in Ghana, there was in 1988–89, as in 1985–87, a marked bi-seasonality for the ali, poli and watsa nets reaching as far as Vridi in 1989).
For the Ghanaian beach-seine, the years 1987–89 were revolutionary for the two sardinella species since they showed a total de-seasonalization of catches, with medium values each month. This phenomenon was doubled for S. maderensis, by delocalization throughout Ghana; the traditional seasonality for this species was, however, more marked in the two western sectors. On the contrary, the anchovy regained the seasonal rhythm of 1982–84 which was disrupted in 1985–86.
The mackerel re-appeared and showed a high abundance in all sectors of Côte d'Ivoire-Ghana in the major cold season.
The length frequencies of the catches of S. aurita and S. maderensis were available in fine strata by month and gear for 1988, 1989 and 1990. These permitted for the first time an analysis of their spatia-temporal distribution in the entire upwelling area.
The spatial strata retained were the ten sectors from Benin to Côte d'Ivoire (Figure 1). The temporary strata were the months, the gears were those of the artisanal and industrial fisheries. Data were required from research vessels which operated in the area. The length frequencies examined were raw, as in the case of Benin, Togo and Ghana for S. aurita and for the data on S. maderensis for Côte d'Ivoire. It would have been better to have these frequencies extrapolated over the total catches, as was the case with the S. aurita data for Côte d'Ivoire.
The lengths caught were evidently affected by the selectivity of fishing gears. The small-meshed gear, the beach-seine and poli, could catch sardinellas as small as 6 cm, but sometimes also caught larger ones. The watsa and ali nets, which have large meshes, generally catch large fish of lengths 20 cm or more. The set net caught only large-sized fish.
The method used involved examining each elementary distribution of length frequencies, in separate/different gears, noting where the modes were, and tabulating these by month and sector (See Table 11 for S. aurita and Table 12 for S. maderensis). (See also Figures 12a–12d which give catches of round sardinella by size class and by gear).
The Case of Sardinella aurita
A large variety of modal lengths ranging from 6 to and 23 cm are constantly observed. To evidence the general characteristics of the distribution, the lengths were regrouped into the following three categories:
Small, length 10 cm or less, considered as class 0+
Medium, length 11–14 cm
Large, length More than 15 cm
The small round sardinellas were caught solely by the artisanal gears, mostly the beach seine. In Côte d'Ivoire, constant catches by beach-seine were noted of very small individuals (6–10 cm long) in sector 7 from mid-1989 to mid-1990. The sizes of sardinellas caught by the beach-seine in the Ivorian sector 6 (Jacqueville) were again unknown. In Ghana, sporadic catches of small-sized S. aurita by the poli in the Greater Accra sector were observed. It is possible that the beach-seines which operate mostly in the Volta sector where they take a large quantities of small sizes anchovies, also catch the small-size sardinellas, but this has never been confirmed. Catches of very small-sized S. aurita in the beach-seines of Benin and Togo are also noted.
Medium-size S. aurita were caught in Côte d'Ivoire by the watsa nets of Vridi in sector 4 only and in all seasons. Catches of medium-sized sardinellas were hardly seen elsewhere along the Ivorian coast. In Ghana, the ali, poli and watsa nets operating in the Greater Accra sector caught medium-sized sardinellas mostly during the major cold season. The semi-industrial vessels also sometimes caught such sizes from this sector. Finally, some catches in Togo were noted. The regrettable absence of length frequency data in the western and central sectors of Ghana since the good catches of the major cold season of 1988 and beginning of 1989 did not permit any possible conclusion regarding the fishery of medium-sized round sardinellas in these sectors. Except for this reservation, it was concluded that the medium-size round sardinella were remarkably rare in the catches throughout the area as a whole.
The large-size round sardinella constituted the bulk of the catch (in weight) for the whole area under consideration. In Côte d'Ivoire, the industrial purse-seiners caught these in all sectors and in all months. The watsa nets also catch them in sector 4 in all seasons and only seasonally during 1988 and 1989 in sector 7. Also noted was the quasi-absence of individuals of less than 18 cm length in the industrial purse-seiners catches whatever the sector. On the other hand, modal lengths of over 20 cm were often observed in their catches. Such large-size fish were present particularly in sector 7. These were found, to a lesser degree, in the catches of the watsa nets. These very large S. aurita were the only ones constantly found throughout the area under consideration. In the Greater Accra sector of Ghana the modal lengths of the large-size S. aurita observed were systematically smaller than those in Côte d'Ivoire, and never larger than 20 cm. The lengths of the fish caught by the artisanal and semi-industrial gears were similar and were in the 16–20 cm range. The length samples which first became available in the Western and Central sectors of Ghana at the end of 1989 showed catches of constant lengths in the 17–18 cm range. It was noted that in Benin catches of very large S. aurita were made in the major cold season (probably the effect of the sardinella ring net selecting only large-sized fish).
In conclusion, the synoptic examination of the length frequency of S. aurita showed notable heterogeneities:
In this way, the two sectors of the west of Côte d'Ivoire; were clearly contrasted with the eastern sectors of Côte d'Ivoire and of Ghana.
Recruitment is observed over a long period in the extreme west of Côte d'Ivoire, and there is probably another recruitment in the east of Ghana. This does not eliminate the possibility of recruitment in the other sectors.
The absence of length data extrapolated for all months and in all sectors (this is the ideal situation) allowed no conclusion to be drawn on possible migration or of simple modeling of the life cycle in time and space as desired. It is obviously necessary to continue to amplify the collection of length data from the whole coastal area as the catch figures are simultaneously collected and to which can be applied these length samples for reliable extrapolation.
5.4.1 Sardinella aurita
In Ghana, the beach seine which exploits mainly the juveniles shows the principal nursery zones to be situated in the east of Ghana in the Volta Region. The periods of maximum abundance correspond to the second part of the major cold season and, rather irregularly, to the minor cold season. There is a slight delay with regard to the period of maximum abundance to the ali, poli and watsa nets which exploit all the age-classes in the three sectors west of Ghana.
In Côte d'Ivoire, the periods of abundance of the industrial and artisanal fisheries corresponded perfectly during the two cold seasons. In the two central sectors (5 and 6) where the artisanal fishery was less active, the purse seiners exploited only fish with modal sizes of 18–23 cm. In the two adjacent sectors the artisanal fisheries exploited individuals with modal lengths of 11–20 cm, even though the industrial fishery selected the very large fish.
The distribution of the abundance index in the whole region evolved between 1982 and 1989 as follows:
In 1982 and 1983, the maximum abundances were observed throughout the region (except in the Volta Region) solely in the major cold season with a slight displacement in appearance from east to west. This corresponds to the east-west propagation of the upwelling.
In 1984, the upwelling was weak in Côte d'Ivoire and there was a higher index of abundance than in Ghana.
the years 1985-1987 were characterized by strong upwellings, notably the intense minor cold season mostly in the west of Côte d'Ivoire, and the resource seemed to react in three ways to this situation: (i) there was an increase in the modal lengths exploited (along the entire coastline), (ii) the appearance of two peaks of abundance each year was established; that is in the minor and the major upwelling seasons and (iii) the resource appeared to be displaced towards the western parts of Côte d'Ivoire and Ghana.
in 1988 and 1989, the upwellings were fairly weak; the pattern in 1988 was similar to that of 1982–83, and 1989 resembled 1984 (decline of abundance in the east of Côte d'Ivoire and in Ghana). There must be pointed out, however, contrary to the difference in the years 1982–1984, the sustained marked abundance in the minor cold season.
5.4.2 Sardinella maderensis
Generally, it appeared that this species occupied the ecological niches left available by the S. aurita. Thus, the juveniles were exploited by the beach-seine in Ghana essentially from November to May (in the warm season) in the same region as the S. aurita. For the other gears, it was difficult to identify the period of abundance of S. maderensis, for this species appeared in the warm season and in the cold season in a fairly irregular manner. One of the explanations of this irregularity could be given by the observation in Côte d'Ivoire of the nearly perfect complementarity in the same site of the periods of abundance for the artisanal and industrial fisheries. The offshore-inshore migration resulting in problems of accessibility to different types of fisheries could also be involved.
It was noted that, overall the periods of abundance of S. maderensis show a complementarity in space and time with regard to those of S. aurita. This was valid in the years other than 1985–87 during which the favourable oceanographic conditions permitted, a better cohabitation of the two species.
The study of the structure of CPUE for the whole region permitted the definition of two apparent stocks: one to the west of Côte d'Ivoire (sector 7 and to a small extent sector 6) and the other in Ghana extending to the east of Côte d'Ivoire (sector 4). The Ivorian sector 5 did not show any periods of great abundance. This would be confirmed by the oceanographic survey data. In the same way, a weak abundance of the species between Ghana and Benin was noted of which only the sardinella nets of Benin caught some large individuals which could have originated from the central stock towards Nigeria.
5.4.3 Engraulis encrasicolus
This species, exploited essentially in Ghana and Togo, showed characteristics similar to those of S. aurita in its spatia-temporal distribution. Periods of high abundance of the species in the beach-seine in the Volta region of Ghana showed a slight delay with respect to that of the S. aurita in the major cold season. The catches of the ali, poli and watsa nets were principally made in the two central sectors of Ghana as for the round sardinella, but the maximum catch of anchovies was made in these two sectors where the catches of round sardinella were minimal. This observation is in line with the opinion of the Ghanaian scientists, according to whom the anchovy is a target species only during periods of low abundance of the round sardinella.
In 1985 and 1986, the anchovy disappeared from the catches of ali-poli-watsa nets in all the sectors of Ghana even though the catches beach-seine were intensified during the same time in the Volta Region. It is difficult to understand the cause, especially since the abundance of the round sardinella in the central sectors was low in 1986.
5.4.4 Scomber japonicus
The CPUE of the chub mackerel followed an erratic pattern which is difficult to describe. On the contrary, it is interesting to draw comparisons among the different areas as well as in relation with the other species.
The mackerel appeared in simultaneously (same months) in all areas of Côte d'Ivoire and Ghana in 1986, 1988 and 1989. It generally appeared to stay much longer in Côte d'Ivoire than in Ghana.
The high CPUE of mackerel corresponded generally to the notable CPUE of the round sardinella but did not seem to have any connection with those of the flat sardinella. This was the case for all years except 1989.
Mackerel appeared in the catches only during the major cold season and never for more than three months.