7 SURVEYS IN THE ATLANTIC OCEAN OFF SOUTHWEST AFRICA (Contd.)

7.2 ANGOLA, 1985–92

Angola, located within the zone of enhanced biological production of the southwest African shelf represented an appropriate choice for a DR. FRIDTJOF NANSEN survey programme. Following discussions between the Angolan Government, FAO and NORAD in 1984 a series of surveys were planned for 1985. The objectives of the programme were established between the Instituto de Investigação Pesqueira, Luanda (IIP) and IMR. The 1985 surveys represented the start of a very comprehensive and long-lasting co-operation, which would entail a large effort, by IIP and IMR in the conduct of surveys and in training of scientists, but which also involved active participation of FAO in support of fishery research in Angola.

7.2.1 History of Angola's fisheries

Angola has an important history of fisheries. A considerable fishmeal industry supplied by purse-seine fleets was developed in the early 1950s with total catches exceeding 300,000 t by the middle of that decade (Campos Rosado, 1974a). After a decline around 1960 the production increased and reached nearly 600,000 t in 1972, but then the fishery collapsed during the war for independence. According to Campos Rosado (1974a) the capacities of fishmeal plants and the purse-seiners were about equal on the Lobito coast, (ICSEAF Division 1.2) and the southern, Namibe to Cunene, coast (within ICSEAF Division 1.3), where the latter had higher landings. Pilchard (Sardinops ocellatus) dominated the high landings in the south in the mid 1950s, but this species declined in the early 1960s and was replaced by horse mackerels (Trachurus capensis and T. trecae). Further north annual catches of sardinellas (Sardinella aurita and S. maderensis) reached about 150,000 t (1969) and of horse mackerels 180,000 t (1966). The existence of a high potential of small pelagic fish was further confirmed by the yield of sardinellas and horse mackerel in a licensed USSR industrial fishery in Angola from the late 1970s to the late 1980s.

It must be assumed that pilchard or sardine (Sardinops ocellatus) in Angola represents a transboundary extension of the Namibian pilchard stock, once a main pelagic resource in the Benguela Current system. The fishery which produced annually more than 1 million t in the late 1960s declined substantially in the first half of the 1970s when the stock collapsed. The total stock biomass which was estimated at about 6 million t in the 1960s had declined to about 100,000 t by 1978. The stock has been monitored from Namibia and it has remained at a level of some hundred thousand tonnes since the collapse.

The depletion of the main Namibian stock must have affected the availability of pilchard in Angolan waters. Pilchard formed part of the raw material for the Angolan fish meal industry which was established in the 1950s and worked until 1973. Campos Rosado (1974a) provides back data on the purse-seine landings for this industry and their species composition. These data show annual landings of pilchard in Baia dos Tigres and Tombua up to about 150,000 t in the late 1950s followed by a period of very low landings until a recovery in 1969 to about 60,000 t and with a mean for the years 1969 to the apparent end of the fishery in 1973 of 42,000 t. These data may include some sardinellas as they were not distinguished in the landing statistics.

Table 7.24 History of the sardinella fishery in Angola (t)

Table 7.24 summarizes historical data on sardinella landings from the area. The fishery is reported mainly from ICSEAF Divisions 1.1 and 1.2 which roughly correspond to the regions Luanda-Cabinda and Benguela-Luanda respectively. Prior to 1973, annual landings did not exceed 160,00 t and were apparently at a very low level from 1974 to 1976 when fishing was resumed by licensed USSR factory expeditions. Total landings exceeded 200,000 t in the late 1970s and in the mid 1980s, but then, at least partly as a result of management measures, declined to about 100,000 t by the end of the 1980s. Effort data from the fishery are not available, but a comparison of catch levels in the mid 1980s with the estimates of biomass from the surveys indicates very substantial rates of fishing mortalities even when allowing for some underestimation of biomass with the acoustic method. The observed trend of decline of biomass from 1985 to 1989 also supports the impression of a stock affected by the fishery. The reduced catches in recent years may have caused some stock recovery.

As reported by Campos Rosado (1974a) horse mackerels were principal species of the Angolan purse-seine fishery up to 1973. In ICSEAF Division 1.3 landings ranged between 100,000 and 200,000 t in the period 1964–73, but this represents a mixture of unknown proportions of the two species. Landings of Cunene horse mackerel (T. trecae) in Division 1.2 reached nearly 50,000 t as a mean of 1956–59, but then declined with only a few years exceeding 10,000 t up to 1973.

Reported landings for both species in ICSEAF Sub-Area 1 for 1973–1986 and the FAO Area 47 for 1987–1988 are shown in Table 7.25. The ICSEAF data showed that the main part of Cape horse mackerel (T. capensis) landings derived from Divisions 1.3 and 1.4, while the Cunene horse mackerel was reported from Divisions 1.1–1.3. There are uncertainties regarding the proper identification of these species in the landings particularly as regards the high catches of Cunene horse mackerel in the period 1973–79.

Table 7.25 History of horse mackerel landings (1,000 t)

Over the period 1976–89 the stock of Cape horse mackerel yielded an annual catch of between 500,000 and 700,000 t with no trend of decline.

For the southern sub-stock of the Cunene horse mackerel there was no decline of nominal stock size and it seems unlikely that the rate of exploitation has been substantial. For the stock north of Benguela where the reported landings exceeded 50,000 t in 1986 the impact of the fishery on the adult stock may have been significant and may have caused the observed decline of the stock size.

Campos Rosado (1974b) also has reported on the deep-sea shrimp fishery by Spanish trawlers prior to 1972. Landings increased from about 1,200 t in 1967 to about 8,100 t in 1972 and was reported as a mixture of the three species: deep-sea rose, striped red and scarlet shrimp caught between 200 and 800 m depth. A licensed fishery was resumed in 1985 with catches of about 9,000 t/year until 1988 when, partly following management restrictions, they declined to about 5,500 t with a further decline to about 4,000 t in 1992. Of this roughly one-third was reported as striped red shrimp, the rest being rose shrimp (FAO Yearbook of Fishery Stat. Vol. 74). Data on effort in this fishery is not available, but there is a close correspondence between the decline in the estimated indices of abundance and in that of the landings over this period. A similar decline was observed over the same period in the density of the Benguela hake which forms a by-catch in the shrimp fishery. The state of the stock of the hake in the shrimp fishing areas could thus perhaps be used as a guide to the conditions of the shrimp stocks as the condition of the hake is easier to monitor than that of the shrimp stocks.

7.2.2 Surveys with the DR. FRIDTJOF NANSEN, 1985–93

Survey objectives and effort

Against this background the principal objectives of the survey programme were:

1. Investigations of the marine fish resources of Angola with emphasis on the distribution, composition and abundance of the stocks of small pelagic schooling fish.

2. Investigations of the stocks of demersal fish on the northern shelf by bottom trawling using the swept-area method.

3. Investigations of the deep-sea shrimp and hake resources on the slope north of Luanda was (included in late 1985).

In later surveys the swept-area programme for the assessment of demersal fish was extended to cover also the regions Benguela-Luanda and Cunene-Tombua. More detailed objectives included the collection of data on size, sex and maturity of the main species for biological studies, collection of taxonomic material for the preparation of a national species guide and observations on the type of bottom from echograms.

The swept-area method was applied with standardized trawl hauls semi-randomly positioned assuming effective fishing over the 18.5 m distance between the wing tips.

The FAO Species Identification Sheets for Fishery Purposes, Fishing Area 34, and FAO World Species Catalogues were used as the main tools for taxonomic identification.

With the assistance from the Nansen programme, FAO undertook the preparation of a field guide for identification of Angolan marine species of economic interest (Bianchi, 1986).

A programme for the monitoring of the main hydrographical features off the Angolan coast was followed from the start with 5–6 fixed main profiles and a more detailed coverage around the estuary of the Congo River.

The 1985 survey programme included four complete coverages of the Angolan shelf and slope from the Cunene River in the south to Cabinda in the north distributed over the year in order to cover the different seasons. An extension was later made to allow two further coverages during the first half of 1986. The objectives for this extension were to study possible inter-annual variations of the distribution and abundance of the stocks of small pelagics and to intensify the trawling programme in order to obtain a better database for the assessments of the demersal resources. An additional objective, time permitting, was to explore offshore parts of the Benguela Current system for the occurrence of horse mackerel, a possibility suggested by the distributional characteristics of a closely related species off the Peruvian coast.

In January 1989 a new survey programme was agreed upon with largely the same objectives as that of the 1985/86 surveys. Special interest was expressed in obtaining new assessments of the important pelagic resources sardinellas and horse mackerels. In 1990 Angola requested an extension of the programme with two surveys in 1991, and a further survey was added in 1992. After the start, in 1990, of a programme of surveys in Namibia, the surveys in Angola, especially in the south, attained increased importance since they covered the northern part of the Benguela Current system. Conversely the surveys off northern Namibia provided information on stocks shared between the two countries.

Also the shelf north of Angola, off Congo and Gabon was covered by a number of surveys programmes starting in 1985 and repeated in 1989 (see Section 7.3).

Operational data for the surveys

Table 7.26 shows the details of the surveys. The degree of coverage is somewhat higher than usual, because a certain amount of steaming not related to acoustic surveys has been included, however, the level indicates a high degree of coverage.

Table 7.26 Operational details of the Angolan surveys 1985–92

All surveys covered the shelf area from a depth of about 20 m along the shore out to the shelf edge at about 200 m depth. The slope down to 500 m depth was also covered extensively, while the five or six hydrographic sections, which included acoustic recording, extended a further 20–30 nmi into the ocean. As an example, Figure 7.16 shows the cruise tracks with fishing and hydrographic stations for one of the surveys (VII).

Figure 7.16 Angola: Example of coverage during a survey. Course tracks of survey 1/89 in the three regions Cunene-Benguela, Benguela-Luanda and Luanda-Cabinda

During survey V (1/86) a few exploratory tracks were made outside the shelf, extending up to 240 nmi offshore in search for possible resources of horse mackerel in these oceanic parts of the Benguela Current, but only squid was observed. Except for some concentrations during survey IV (4/85) of medium sized horse mackerel 2–15 nmi outside the shelf edge off Baia dos Tigres, very little pelagic fish was observed at depths of more than about 200 m, and the main concentrations were located over the shelf inside this depth. However, the possibility that offshore resources of small pelagic fish may have been missed by the programme cannot be excluded.

Acoustic equipment

The EK500 was used in 1991 and 1992, replacing the EK400 system used in 1985–89. As discussed in Chapter 6, signal saturation could occur in the EK400 system under special circumstances, dense schools close to the transducer. This technical problem was solved by the new system. Schools of adult small pelagic fish may have a specially high density and time-series of estimates of their biomass may thus be biased. A comparison of the frequency distribution of integrator readings (per nmi averaged over 5 nmi) from the two systems may indicate whether saturation has occurred. The effect would be to suppress high values. Table 7.27 shows such frequencies by levels of integrator values for sardinellas between Cabinda and Luanda, comparing survey 2/86 and 2/92 with biomass estimates of 130,000 t and 154,000 t respectively and for horse mackerel on the Baia dos Tigres shelf comparing surveys 4/85 and 1/91 with estimates of 370,000 and 410,000 t respectively. There was a somewhat higher proportion of high integrator readings in the 1991 surveys which may indicate that saturation took place in the EK400 system. A quantification is not possible, but the possible bias of time series can not be very substantial.

Table 7.27 Comparison of frequency distributions of integrator readings between surveys with EK400 and EK500 systems (Unit 0.1 * m²/nmi²)

Continental shelf and slope

Angola has a coastline of 800 nmi of which that of Cabinda is 45 nmi. The total shelf area is about 17,000 nmi² and the mean width is thus about 20 nmi. Three main regions are distinguished:

Cunene River to Benguela (sub-divided into Cunene-Tombua and Tombua-Benguela) Benguela to Luanda (also identified as Ponta das Palmeirinhas) Luanda to Cabinda.

Estimates of the extension of the shelf by depth ranges in the different regions are shown in Table 7.28. These are based on measurements taken from special bathymetric charts prepared from echosounding during the surveys. The southernmost sub-region from Cunene up to Tombua has a relatively wide and shallow shelf, but with a steep slope except in the extreme southern part. From Tombua to Benguela the shelf is very narrow and the slope too steep for any bottom water trawling. From Benguela northwards the shelf is generally wide all the way up to Cabinda, but with a narrow part off Luanda and a steep “canyon” outside the Congo River. The slope down to 600 m and beyond has a gradient which over wide distances is sufficiently low and smooth to permit bottom trawling.

From five of the surveys, records are available of the character of the bottom with reference to its suitability for trawling based on analysis of the echograms. These observations are set out in Figure 7.17 and show that most of the shelf and the slope can be fished with trawls. Some areas may be too rough especially along the narrow shelf from Tombua to Benguela and off N'Zeto. In the northern areas offshore oil exploration and production may prevent or endanger trawling operations, not only near the installations, but also in areas of previous exploration and along supply vessel routes where debris may have been dumped.

Table 7.28 The shelf areas off the Angola coast by depth ranges and regions (nmi²), and length of coastline by region (nmi)

Figure 7.17 Angola: Sea bottom types by area Cunene River-Benguela, Benguela-Luanda and Luanda-Cabinda

Hydrography¹

The literature describes the main circulation off the Angolan coast as being dominated by the Angola Current flowing mainly southward and apparently formed by the southeast branch of the South Equatorial Countercurrent and southward-turning waters from the north branch of the Benguela Current.

The Angola Current is a generally southward moving stream from 9° (Ponta das Palmeirinhas) to 16°S (Tombua) that penetrates to depths of 250–300 m and covers not only the shelf regions, but also the continental slope, with surface velocities exceeding 50 cm/sec (about 1 knot).

The Benguela Current, part of the anticyclonic sub-tropical gyre of the South Atlantic and forming its eastern boundary, flows parallel to the southwest African coast in a north to northwest direction. Streams of the Benguela Current proceed northward along the shore and reach 13°–14°S (Cabo de Santa Marta), and they form a divergence zone along 11°E from 17°S to 13°S, as a result of an interaction with the Angola Current.

The series of simple oceanographic data acquired during the DR. FRIDTJOF NANSEN surveys made it possible to add details to this general large-scale picture, particularly as regards the regime along the coast and over the shelf. Confirming previous observations the data collected thus indicate that the direction of the surface current over the shelf area is southward only in the (southern) summer season. During most of the autumn, winter and spring seasons a surface current flows northward along the coast causing inshore upwelling in various locations. That the flow is towards the south in summer is demonstrated by low salinity surface water originating from the Congo discharges being found off N'Zeto and Punta das Palmeirinhas and remnants even off Lobito.

The offshore waters were not covered by the surveys, but judging from the observations in the frontal area between the Benguela and the Angola Current at about 16°S it seems likely that the more offshore waters off central Angola are also in winter affected by the southward turning gyre formed by a right-hand branch of the Benguela Current and the southeast branch of the South Equatorial Countercurrent.

The characteristics of the hydrographical regimes along the coast and their ecological significance can be summarized as follows:

In the region from Benguela to Cabinda the hydrographical regime of the shelf waters is characterized by a marked seasonal shift. According to the observations during the surveys the surface current flows southwards during less than half the year, in the period January to May. The sea surface temperature at this time is high: 27°–28°C and the surface layer is stable with a well developed thermocline at 30–50 m depth. The temperature below the thermocline decreases from about 20°C at 50 m to 15°C at 200 m and the oxygen level declines to 1.5–2 ml/l at 200 m. The salinity of the top surface layer is reduced by mixing with the outflow of the Congo River, but this is unlikely to have effects on the fish fauna except for the area around the river mouth, although turbidity and bottom deposits are noticeable over large areas.

¹ This section is a brief summary of a thesis prepared on the basis of data from the 1985/86 surveys by Antonio Fontes Pereira, IIP, Luanda. His contribution is gratefully acknowledged.

In late May-June the change of season initiates with a reduction in the surface temperature apparently starting in the north and with a reversal of the direction of the coastal current. Extreme conditions of the winter season occur in August-September with surface temperatures of 20°–22°C, some times as low as 19°C inshore. Upwelling results in pockets of cold water at locations along the coast especially outside and north of cape configurations along the coastline. The discharges of the Congo River now turn northwest along the coast past Cabinda.

With the approaching spring season in November-December the surface layer heats up with temperatures reaching 25°–26°C, the current still flows predominantly northward, but at reduced velocity and upwelling only occurs in isolated spots.

The most significant feature of this regime in an ecological context is the upwelling process which starts in May-June, has its maximum in August-September and probably ceases towards the end of the year. The upwelling results in intensified primary production of high significance for fish production and distribution. Especially the small pelagic fish, sardinellas and horse mackerel and juveniles of other species low in the food chain, are likely to be affected by these phenomena. The high rate of production in upwelling locations sometimes results in areas of surface water being discoloured by high plankton content and drifting northwards along the coast. Such areas were found in late August 1985 and one area off N'Zeto coincided with a dense distribution of surface schooling sardinella. It seems likely that the upwelling along the coast has a higher persistence and intensity in certain locations such as the area between Punta das Palmeirinhas and Luanda. This may affect the distribution of the plankton-consuming types of fish.

The pronounced seasonal changes in oceanographic parameters were found to occur only in the upper water masses over the shelf. Below the discontinuity layer, which in summer reached down to 60–70 m depth the seasonal change consisted only in 1°–2°C higher temperature in summer down to about 100 m and in that water with oxygen levels below 1 ml/l occured shallower than 100 m. In the 200–500 m depth range, the environment of deep-sea shrimp and Benguela hake stocks, conditions are quite stable throughout the year with temperatures ranging from about 14°C at 200 m to 7°–8°C at 500 m and oxygen levels of around 1 ml/l. From Punta das Palmeirinhas southwards, there is a minimum below this level at 300–400 m depth.

Since fish often adjust their reproduction strategies to current systems and to cycles in basic productivity it seems likely that the spawning seasons and areas of the main Angolan species have seasonalities which coincide with the observed oceanographic patterns.

It should be noted that although marked seasonal changes only affect the upper 100 m or so of the water column, the high productivity here is spread out and available over wider depth ranges. Predating fish undertake vertical migrations and some of the plankton consumers, e.g., mesopelagic fish have regular diurnal migrations between the surface layers and depths of several hundred metres.

Strong seasonal variations are also a characteristic of the regime in the region from Cunene to Namibe on the south coast. Here the variations consisted mostly in a displacement in a north-south direction of the frontal zone between water masses from the Benguela and the Angola Currents. In summer (February) the front had moved towards the southern part of the area. The intrusion of Angolan water from the north was demonstrated by a well developed surface tongue of warm water stretching south in the open water some 20 nmi outside the shelf edge. Cooler water then extended northwards inshore with indications of upwelling even in the summer. With the displacement northwards of the front in autumn and winter the cold Benguela water covered the shelf up to Punta Albina and the distribution of isolines indicated strong upwelling.

The ecological significance of this regime lies first of all in the varying influence of the Benguela Current system. The characteristic fauna of this system follows its displacements, but perhaps with some lag in the southward return migration. Seasonal and interannual variations in the position of the frontal area are then likely to result in changes in the availability to Angola of the stocks shared with Namibia.

The upwelling process in this area is of great significance for its productivity also in respect of the true Angolan fauna such as sardinellas and Cunene horse mackerel which predominate in the northern part of the area. The hydrographical findings indicated that the upwelling which causes enhanced primary productivity may occur not only along the Baia dos Tigres shelf, but also during a considerable part of the year, further north along the narrower shelf from Tombua to Namibe. In an ecological context it may therefore prove appropriate to consider this part of the coast as being included in the southern regime affected by the Benguela Current conditions of low temperature and high productivity.

Reporting

Cruise Reports, containing brief descriptions of the work done and the main findings, were issued upon the completion of each cruise (IMR 1985a, 1985c, 1985e, 1986a, 1986d, 1986e, 1989b, 1989c, 1989g, 1991b, 1991c, 1991e (in 1991d), 1992f.)

Furthermore, two Summary Reports were produced (Strømme and Sætersdal, 1986 and 1991).

In all these reports the results were presented by the three main regions identified above.

Three main regions are distinguished:

Cunene River to Benguela (sub-divided into Cunene-Tombua and Tombua-Benguela) Benguela to Luanda (also identified as Ponta das Palmeirinhas) Luanda to Cabinda.

These regions do not correspond directly to areas which can be distinguished on an ecological or faunistical basis, but contain some differences in these respects.

Cunene-Benguela region

The southernmost part of this region, the relatively broad shelf from Cunene to Tombua represents the frontal zone between the Benguela Current and the Angola Current and hence also of the stocks which belong to the Benguela ecosystem: pilchard, Cape horse mackerel and Cape hake. All these species seem to undertake migrations in the border area between Angola and Namibia.

In many surveys the distribution of some of the pelagic resources was found to extend from Cunene to a little northwards of the Baia dos Tigres shelf up along the narrower shelf towards Namibe. It is likely that the species found in this continuous distribution, Cunene horse mackerel and sardinellas, constitute sub-stocks which are more or less independent from the stocks of these fish found north of Benguela, since only a few and usually scattered formations were located on any of the surveys over the long and narrow stretch of shelf from Namibe to Benguela.

Benguela-Luanda region

This region lies within the Angola Current regime with high surface temperature and a well developed thermocline in the summer, but with a northward flowing surface current in winter and spring at which time the thermocline largely breaks down and inshore upwelling occurs locally. The locations of upwelling seem to affect the distribution of small pelagic fish. There is no ecological barrier towards the region further north and it seems probable that pelagic fish migrates between the two regions and that other forms of mixing may occur.

Luanda-Cabinda region

This area can ecologically be considered as an extension of the Benguela-Luanda shelf, but the northern part and in particular the shelf off Cabinda is more directly affected by the discharges from the Congo River.

7.2.3 Pelagic resources

The acoustic integration system provided the observations of fish density from which distribution charts were developed. On the basis of identification by trawl sampling and fish behaviour two groups of pelagic fish were distinguished:

The survey results will be presented by species or species groups, following the list given above. For several species the results will also be presented by regions.

Table 7.29 shows the biomass estimates for type 1 pelagic fish, which consist mainly of sardinellas, but with small amounts of anchovy and ilisha from Benguela northwards.

Table 7.29 Angola: Estimates of biomass of pelagic fish type 1, mainly sardinellas, by regions and surveys (1,000 t)

There appeared to be some change with time in the mean estimates. For the main regions of distribution of type 1 pelagic fish, from Benguela to Cabinda, there was a decline from a mean of 310,000 t for the four 1985 surveys, to 255,000 t for the two 1986 surveys and down to 195,000 t for the three 1989 surveys, but then the estimates increased to 260,000 t in 1991 and 280,000 t in 1992.

Figure 7.18 Cunene-Namibe: Distribution of type 1 pelagics mainly pilchard in surveys 1/85 (left) and 1/91 (right)

Figure 7.18 shows distribution of the type 1 species in surveys 1/85 and surveys 1/91, mainly pilchard which may serve as examples of the characteristics of the distribution of this type of fish. The whole series of data shows that aggregations in school areas were usually of limited extension, not exceeding 10–20 nmi along the cost, always on the inner part of the shelf, and often quite close to the shore. When pilchard was located it was found on the Baia dos Tigres shelf, often inside the bay itself.

Pilchard or Sardine

Pilchard was found in nine out of the twelve surveys.

Table 7.30 Cunene-Tombua: Biomass estimate of pilchard by surveys (1,000 t)

Table 7.30 shows biomass estimates of the pilchard by surveys. The surveys with the highest biomass were all in late winter, August-September, which coincides with the northern position of the oceanographic front indicating seasonal migrations of parts of a joint Namibian-Angolan stock.

Pilchard was in all cases located inshore on the extreme southern part of the Angolan shelf from Baia dos Tigres southwards. The 12 estimates of the biomass of the species ranged from 0 to 210,000 t with an average of about 50,000 t. This average is not very meaningful since there is a seasonal migration pattern with the highest abundance in Angola in late winter-spring when the Benguela/Angola front is in its northernmost position.

Sardinellas

The two species round sardinella (Sardinella aurita) and flat sardinella (S. maderensis) occur along the whole coast from Cunene to Cabinda and with an apparently extended continuous distribution along the coasts of Congo and Gabon.

The relative proportion of the two species was assessed from their incidence and catch rates in trawl samples from the 1985–89 surveys. These data indicated that in the southern area, Cunene-Benguela, the round sardinella was more common than the flat sardinella. In the samples from Benguela northwards the overall mean catch rates were similar, but with a slightly higher incidence for the flat sardinella. In the following analysis the two species were dealt with together.

Table 7.31 Sardinellas: All surveys 1985–1992. Mean biomass by regions and densities over the shelf area, 0–200 m and by 100 nmi of coastline

Table 7.31 shows the estimated biomass of sardinellas (with small amounts of anchovy and ilisha) as means for all 12 surveys by regions and the densities by unit shelf area and by 100 nmi of coastline. By shelf area the density was highest in the Benguela-Luanda region, by coastline the densities were nearly the same.

Cunene-Benguela

Sardinellas were found in small school areas inshore, mostly south of Namibe, but in a few surveys also in the north close to Benguela. Table 7.32 shows that the estimated biomass of these species in most surveys ranged between 0 and 40,000 t, but reached 110,000 t in April 1985. Both species, round sardinella (Sardinella aurita) and flat sardinella (Sardinella maderensis) occurred in the area.

No sardinellas were found in the 1991–1992 surveys in this area.

Table 7.32 Cunene-Benguela: estimates of biomass of sardinellas, including small amounts of anchovies (1,000 t)

Benguela-Luanda

Of the pelagic type 1 fish, the two sardinella species, round sardinella (Sardinella aurita) and flat sardinella (S. maderensis) were by far the most common. Anchovy (Engraulis encrasicolus) and West African ilisha (Ilisha africana) were found in minor aggregations inshore, but the distribution charts will mainly represent the flat and round sardinellas. Figure 7.19 shows characteristic distributions of this group of fish from surveys 2/85 and 3/89. The sardinellas were found in more or less well defined school areas located inshore in most surveys near Benguela-Lobito and otherwise often near the capes along the coast, Cabeça da Baleia, Punta do Morro, Cabo São Braz and Punta das Palmeirinhas. There seem to be less high density areas during the autumn surveys than at other times of the year, possibly an effect of the fluctuations in oceanography and production in the surface layers.

Figure 7.19 Benguela-Luanda: Distribution of sardinellas in surveys 2/85 (left) and 3/89 (right)

Table 7.33 Benguela-Luanda: Incidence and mean catch rates of the two sardinella species, 1985–89

Because of the low catchability of the sardinellas, the data on the catches are of little interest for studies of changes in abundance, but may serve for a comparison of the two species. Table 7.33 gives the percentage incidence and the mean catch rates in hauls with catch of the species within the shelf, 0–200 m for the period 1985–89. There is some similarity in the trends of the figures for the two species. The two first surveys show relatively high levels followed by a decline and then an increase in survey 1/86 and survey 3/89. The overall mean incidence is 24% for round and 31% for flat sardinella with mean catch rates of 37 kg and 34 kg respectively. This indicates a slightly higher availability of flat sardinella, but since this species is reported to have a more inshore distribution than the round sardinella, the difference in incidence may have been caused by higher accessibility to the fishing gear.

Table 7.34 shows the biomass estimates of the sardinellas in this region. The levels vary considerably between surveys. There may be a tendency of low biomass in this area in winter-spring.

Table 7.34 Benguela-Luanda: Biomass estimate of sardinellas by surveys and season (1,000 t)

Samples of size compositions of both species showed the presence of modal groups of highly varying sizes from 10 to about 35 cm modal length without any apparent seasonal trend. These incomplete data can probably not be used for studies of growth and recruitment, but they demonstrate that in some years reproduction takes place over wide areas.

Luanda-Cabinda

The distribution of the pelagic fish type 1, consisting mostly of sardinellas is shown in Figure 7.20 exemplified by the situations found in survey 3/85 and 2/86. High density school areas were nearly only found along the inshore shelf between Luanda and N'Zeto and in a few surveys also off Cabinda. In the vicinity of the Congo River mouth the sardinellas were in several surveys found on the outer shelf, some times off the shelf edge as in August 1985.

Figure 7.20 Luanda-Cabinda: Distribution of sardinellas in surveys 3/85 (left) and 2/86 (right)

The mean incidence and catch rates of round and flat sardinella in all hauls on the shelf in the 1985–89 surveys are shown in Table 7.35. The overall means were 20% and 24% respectively with mean catch rates of 37 kg and 39 kg. This indicated a slightly higher availability of flat sardinella, as was the case in the southern area, but this could also be explained by a more inshore distribution of flat sardinella.

Table 7.35 Luanda-Cabinda: Incidence and mean catch rates of the two sardinella species

Table 7.36 shows the biomass estimates of this group of pelagic fish consisting mainly of sardinellas over the Luanda-Cabinda shelf. The mean of all the survey estimates is about 10% lower than that of the Benguela-Luanda region and the mean fish density over the shelf is only about half, perhaps reflecting a lower basic productivity in this northern area.

Table 7.36 Luanda-Cabinda: Acoustic biomass estimates of sardinellas by season and year (1,000 t)

Samples of the size compositions of the two sardinella species showed groups with widely different modal length from below 10 to about 35 cm as in the southern region, but it seems doubtful whether representative sampling was achieved.

Anchovy

An other species of the pelagic fish type 1 group in this area is the anchovy (Engraulis encrasicolus). Anchovy was only found in small and scattered formations on the coast from the Cunene River to the Baia dos Tigres.

Horse mackerels

Southern Angola forms the northern border for the Cape horse mackerel (Trachurus capensis) and also the approximate southern border of the Cunene horse mackerel (Trachurus trecae). The two species are mixed on the fishing grounds in the Cunene-Benguela area and may occur together in the catches. This has caused problems for the identification to species level of the commercial horse mackerel catches in this border region.

Figure 7.21 shows the distribution of horse mackerels in the surveys of April-May and November-December 1985 with the high densities on the outer shelf characteristic of these fish. In some surveys such as in November 1985 and in February 1989 horse mackerels were also found in mid-water up to 5–15 nmi outside the shelf edge. On these occasions the outer limits of distribution appeared to coincide with the tongue of warm Angolan water reaching southwards off the shelf. In many of the surveys there was a more or less continuous distribution of horse mackerel from the Baia dos Tigres shelf northwards past Tombua up to Namibe.

Figure 7.21 Cunene-Namibe: Distribution of horse mackerels in surveys 4/85 (left) and 1/89 (right)

Juvenile horse mackerel was often found in abundance inshore usually T. trecae. Adult horse mackerel had a more offshore distribution and occurred in mid-water in schools and other aggregations of varying density often close to or near the bottom during daytime, but usually lifting and dispersing at night. A bottom trawl can be an efficient gear for these species, as they are close to the bottom during the day.

Table 7.37 Cunene-Benguela: Estimates of the biomass of Cape and Cunene horse mackerels by surveys (1,000 t)

Table 7.37 shows the estimates of biomass of the horse mackerels from the acoustic integration system for each of the surveys. There is a considerable variation in available biomass with a range from 40,000 to 400,000 t. The estimates also show a tendency of a seasonal shift with high levels in the spring and low in the autumn. These are mainly caused by changes in the distribution of the Cape horse mackerel related to the seasonal shift of the front between the Benguela and the Angola Current systems.

The Cape horse mackerel represents the major part of this biomass, but the Cunene horse mackerel is also at times relatively abundant. Their distribution is largely segregated with Cunene horse mackerel dominating in the northern and innermost parts of the Baia dos Tigres shelf and up to Benguela and the Cape horse mackerel the offshore and southern parts.

Cape horse mackerel

Like the pilchard the distribution of this species in Angola is restricted to the Cunene-Tombua region. Its availability depends on migrations out from and into Namibia where this stock which has a total biomass of several million tonnes, has its main area of distribution. Table 7.38 presents an overview of the biomass estimates from the twelve surveys arranged by season. The highest estimates were usually from the last half of the year demonstrating a relationship between the migrations and the northward shift of the border between the Benguela and the Angola Currents. These data indicate that the availability of Cape horse mackerel in Angola exceeded 200,000 t probably over a substantial part of the year in each of the seasons covered by the survey.

Table 7.38 Cunene-Tombua: Acoustic biomass estimates of Cape horse mackerel by seasons (1,000 t)

Cunene horse mackerel

The Cunene horse mackerel is found over the entire Angolan shelf. In the following analysis the biomass data from the autumn surveys have been excluded because they are thought to have been biased. Table 7.39 shows the mean acoustic biomass estimates by regions for the whole survey period and the densities by shelf area and by coastline. The southern region showed by far the highest density both by area and by coastline.

While normally most of the Cunene horse mackerel was found over the northern part of the Baia dos Tigres shelf, its distribution in surveys 1/86 and 3/89 included also the southern part of the shelf where it was found mixed with Cape horse mackerel.

Table 7.39 Cunene horse mackerel: Mean estimates of biomass by regions and densities over shelf area, 0–200 m and by 100 nmi of coastline

These findings might indicate that there is a separate sub-stock of Cunene horse mackerel in southern Angola and northern Namibia which benefits from the high basic productivity in an area that most of the year is influenced by the Benguela Current system.

The relative geographical isolation of this southern distribution area and the fact that both adults and juveniles were found here, the juveniles sometimes in high densities inshore, suggest the existence of a separate southern stock unit benefiting from the high productivity of the Benguela Current system and extending its distribution into northern Namibia.

The more or less continuous distribution observed from Benguela past Ponta das Palmeirinhas and the biological data with similarity of size distributions strongly indicate that there is one main stock over the whole shelf from Benguela to Cabinda, but with a markedly lower density north of Luanda.

Table 7.40 shows time-series of the acoustic biomass estimates over the survey period. There is a high variability especially in the southern region where there is no clear trend. For the Benguela-Cabinda stock there may be a trend of decline from 1985 to 1989 and some recovery to 1992 similar to that found for the sardinellas.

Table 7.40 Cunene horse mackerel: Time series of biomass estimates of by regions, autumn surveys excluded. Estimates from bottom trawl hauls in brackets (1,000 t)

The estimates of the horse mackerel from their part of the bottom trawl catches were as shown in Table 7.40 usually, lower than the acoustic estimates. There was a wide variation in the ratios between the estimates and this may in part have been caused by changes in the size composition of the horse mackerels between surveys.

Other Carangidae, hairtail and barracudas

The other types of pelagic fish associated with horse mackerel have their main distribution north of Benguela. Table 7.41 shows the mean acoustic- and swept-area estimates by regions and for the survey periods from which these data are available. The group “other Carangidae” include bumper, lookdown and false scad. Large-sized false scad were sometimes found in dense schools off Ponta das Palmeirinhas and these concentrations may have been exposed to the industrial purse-seine fishery. This group as well as the barracudas may have declined in abundance, while hairtails were observed in fairly unchanged abundance over the period. Judging from the ratios of the two methods of abundance estimation, these species seem to have a more pronounced mid-water behaviour than the horse mackerels.

Table 7.41 Mean biomass estimates (1,000 t) of groups of type 2 pelagic fish associated with horse mackerels by regions and survey periods. Estimates from random trawl hauls in brackets

On the shelf from Benguela northwards the pelagic fish classified as type 2 consisted of a number of species in addition to the Cunene horse mackerel. Hairtail (Trichiurus lepturus) occurred in abundance over a wide depth range. Among the other Carangidae bumper (Chloroscombrus chrysurus) and lookdown (Selene dorsalis) often appeared in the catches in shallow water together with other species of this family. Barracudas (Sphyraena guachancho) and (S. sphyraena) appeared with low catches, but high incidence. The contribution of the former species was 3–4 times higher than the latter. The Scombridae gave only scarce and small catches with bonito (Sarda sarda) and Spanish mackerel (Scomberomorus trito) occurring more often than chub mackerel (Scomber japonicus) and little tunny (Euthynnus alletteratus). This last species was often observed in schools near the surface in the slope region.

The catchability of these species in mid-water and bottom trawl probably varied considerably. It should be expected to be low for bonito, tuna and chub mackerel and for large-sized schooling horse- and jack mackerels, but high for hairtail, bumper, lookdown, small-sized horse mackerel and probably also barracudas. The catch data will thus only give a rough indication of the true proportions of the various species and when used their probable bias should be kept in mind.

Benguela-Luanda

Time-series of biomass estimates by groups of the pelagic type 2 fish were calculated for part of the surveys using their mean proportions in the catches and the total biomass figures (Table 7.42). In view of the difference in catchability these estimates should be considered as indications only. There may be a trend of declining abundance of horse mackerel.

Table 7.42 Benguela-Luanda: Time series of biomass estimates of horse mackerels and associated species (1,000 t)

Figure 7.22 Benguela-Luanda: Distribution of horse mackerels, other Carangidae and associated species in surveys 1/86 (left) and 1/89 (right)

Figure 7.22 shows the distribution charts of the pelagic fish classified as type 2 between Benguela and Luanda from the surveys 1/86 and 1/89 which provide examples of the typical distribution of these fish. In general they were found in lower densities over wider areas than the sardinellas. The high-density school areas which were usually quite restricted, consisted mostly of large-sized Cunene horse mackerel, false scad (Decapterus rhoncus) or chub mackerel. In a number of the surveys such school areas were located off Punta das Palmeirinhas.

Luanda-Cabinda

Table 7.43 shows the biomass estimates from the acoustic integration system. There was a considerable survey-to-survey variation which may include changes in availability. Thus several of the autumn surveys (April-June) seem particularly low as was also the case with estimates of demersal fish. The mean biomass for all surveys was only about two-thirds of that on the Benguela-Luanda shelf and the mean density of the fish over the shelf area was as for the sardinellas considerably lower, in this case 40% of that in the southern region.

Table 7.43 Luanda-Cabinda: Acoustic biomass estimate of horse mackerel, other Carangidae and associated species by surveys (1,000 t)

Time-series of biomass estimates of the main groups have been calculated applying the mean proportions in the catches for the 1985/86 and the 1989 surveys to the total biomass figures (Table 7.44). The three autumn surveys have been omitted because of suspected bias. There is a declining trend in “Other Carangidae” which demonstrates reduced abundance principally of bumper.

Table 7.44 Luanda-Cabinda: Time series of biomass estimates of horse mackerels and associated species (1,000 t)

Figure 7.23 Luanda-Cabinda: Distribution of Carangidae and associated species in surveys 3/85 (left) and 1/89 (right)

The size samples of horse mackerel were dominated by groups of juvenile or young fish. In most of the surveys the juvenile groups had approximately the same modal size as the dominating group found in the Benguela-Luanda region. This is a strong indication of a unit stock in the two regions. A closer comparison shows that the modal size of these groups tended to be slightly higher in the northern region. This indicates either a higher growth rate in the north, an earlier spawning period or, perhaps more likely, that spawning mainly occurs in the Benguela-Luanda region and that the juveniles in the north derive from the first part of the spawning period and have drifted northwards with the prevailing surface current in winter-spring.

Figure 7.23 shows the distribution of these types of fish exemplified by those found in surveys 3/85 and 1/89. The few high density school areas were mostly located in the south. In most surveys the fish was found distributed thinly over wide areas. There was usually a paucity of fish off the Congo River, perhaps an effect of the discharges.

The proportions between the types of fish in this group were roughly similar to those found on the shelf between Benguela and Luanda with “other Carangidae” and hairtails showing largely similar catch levels as horse mackerel.

Hairtail

Table 7.45 shows the biomass and density data for largehead hairtail (Trichiurus lepturus) this species from the trawl surveys and also the acoustic estimates since this fish is commonly found in mid-water. The bottom trawl probably gave a serious underestimate and it is not unlikely that the true biomass is in excess of 50,000 t.

Table 7.45 Hairtail: Mean biomass estimates based on trawl hauls by regions and survey periods and mean acoustic estimates from the 1985 through 1989 surveys (1,000 t)

Benguela-Luanda

Hairtail was found all over the shelf, but in highest abundance at shallow and intermediate depths as shown by Table 7.46.

Table 7.46 Benguela-Luanda: Incidence and mean catch by bottom trawl of hairtail by depth and surveys

Luanda-Cabinda

Hairtail was found all over the shelf with common incidence rates around 50% in all three depth zones.

7.2.4 Demersal resources

The analysis of the demersal resources is based on the semi-random swept-area stations. The most abundant demersal fish of commercial interest on the Angolan shelf are reviewed below. They include seabreams, grunts, croakers, groupers and Benguela hake.

Table 7.47 shows the biomass estimates for the conventional commercial demersal species: seabreams, croakers, groupers and grunts (big-eye grunt and Benguela hake are not included).

The total 1991–92 estimates ranged from 115,000 to 120,000 t. The stocks seem to have had an increase of abundance in the regions Cunene-Tombua and Benguela-Luanda.

In addition to these demersal fish which have a conventional commercial use in Angola, the surveys identified other groups which have commercial value in other countries and may thus be added to the potential resources for a commercial Angolan fishery. The most important of these are the big-eye grunt with an estimated biomass of 100,000 t, hairtail with 50,000 t and other Carangidae and barracudas with perhaps 30,000–40,000 t. The total biomass of the resources for a mainly demersal fishery seems thus to be of the order of 300,000 t. In addition comes the Benguela hake in the slope with a biomass of about 30,000 t.

Table 7.47 Commercial demersal fish*: Estimates of biomass by surveys and regions (1,000 t)

Cunene-Benguela

The southern part of this area, from the Cunene River to Tombua, has a relative wide shelf (about 1,690 nmi²), with smooth bottom conditions. The northern part, from Tombua to Benguela, has a narrow and rough shelf (about 1,090 nmi²) and only a few hauls were made in this area. No semi-random hauls were made in 1985.

The number of successful semi-random trawl hauls by survey were as follows:

The effort was thus especially low in surveys 2/86 and 3/89 and highest in 1991 and 1992.

As regards species distribution, this region seems to represent a transition zone between the tropical regime in the north and the sub-tropical/temperate regime found further south. The shelf south of Tombua is inhabited by a lower number of species, including the large-eye dentex (Dentex macrophthalmus), the African weakfish (Atractoscion aequidens), the Cape hake (Merluccius capensis) and the Benguela hake (M. polli). Less abundant, but common are the wedge sole (Dicologoglossa cuneata), the John dory (Zeus faber) and the red pandora (Pagellus bellottii).

More sporadic trawl hauls made on the shelf between Tombua and Benguela showed that this part is also inhabited by the large-eye dentex, but other species with affinity for tropical waters are also found, such as the big-eye grunt (Brachydeuterus auritus) and the Angola dentex (Dentex angolensis). Other species are the red pandora, the Canary drum (Umbrina canariensis) and common, but less abundant, the bogue (Boops boops) and the African weakfish.

Table 7.48 shows biomass estimates of the commercially important groups of demersal fish calculated from mean densities by the depth ranges 10–50 m, 50–100 m and 100–200 m.

Among the causes of the variation in the estimates are survey variability and migrations into and out of Namibia. The timing of the surveys has not been favourable for an analysis of possible seasonal variations in the abundance of these demersal stocks. The apparent trend of increase of the last surveys is discussed below.

Table 7.48 Cunene-Tombua: Biomass estimates (1,000 t) of main groups from random trawl survey over the shelf to 200 m depth

Benguela-Luanda

The demersal fauna between Benguela and Luanda was not investigated during the 1985 programme, except for nine test hauls during the November survey. The number of successful swept-area hauls in the subsequent surveys were as follows:

The total shelf area (0–200 m) is about 4,800 nmi² with a mean width of about 18 nmi, the effort of only 15 hauls in survey 3/89 is considered too low and therefore these data will not be used here.

The main species groups in this region were grunts (Pomadasyidae), seabreams (Sparidae) and largehead hairtail (Trichiurus lepturus). Of the grunts, the big-eye grunt (Brachydeuterus auritus) was the dominating species, while three Pomadasys species were far less common. The seabreams were mainly represented by the red pandora, but various Dentex species were also common.

Table 7.49 shows the biomass estimates from the swept-area calculations.

Table 7.49 Benguela-Luanda: Biomass estimates (1,000 t) of main groups from swept-area surveys over the shelf to 200 m depth. Conventional commercial species are summed

The big-eye grunt (Brachydeuterus auritus) and the Cunene horse mackerel (Trachurus trecae) were the most abundant species. Both species as well as the other Carangidae, and the barracudas and hairtails were found to have a more pronounced semi-demersal behaviour than most of the other species caught by the bottom trawl and this may explain a part of the large survey to survey variation in their biomass estimates. Acoustic estimates of the biomass were also made for these groups.

Luanda-Cabinda

The region from Luanda to Cabinda was covered with swept-area hauls during all surveys since January 1985. The number of semi-random hauls were as follows:

The area of this shelf is about 4,800 nmi² and the precision of the biomass estimates from the surveys in 1985 and survey 3/89 is expected to be low.

The main species in the bottom trawl catches were the big-eye grunt, the hairtail, the red pandora, other seabreams (Dentex spp. and Sparus sp.), other grunts (Pomadasys spp.), croakers (Pseudotolithus spp., Miracorvina angolensis, Umbrina spp.) and groupers (Epinephelus spp.), mentioned in their order of dominance.

Biomass estimates of main groups based on mean densities by depth ranges are shown in Table 7.50. The high variation for the semi-demersals, big-eye grunt, horse mackerels, other Carangidae, barracudas and hairtails may in part be caused by changes in their pelagic behaviour from survey to survey. For the demersals this long series of estimates may indicate that they in part depended on the season of observations. If within-year estimates are compared those from April-May and May-June were markedly lower than estimates from August-September and November-December (2/85 compared with 3/85 and 4/85; 2/89 compared with 3/89 and 1/91 compared with 2/91). In the late autumn season April-June the Angola current flows southward over the shelf with Congo water in the surface layers and the thermocline is deep. These conditions may have caused a reduced availability of demersal fish to the bottom trawl. Estimates from this warm season should thus not be used in time series for stock evaluation purposes.

Table 7.50 Luanda-Cabinda: Biomass estimates (1,000 t) of main groups from random trawl survey over the shelf to 200 m depth

The big-eye grunt and the Cunene horse mackerel were also in this region the most abundant species in the bottom trawl catches. The large variations in the estimates for these species and well as for the other semi-demersal: (the other Carangidae, the barracudas and hairtails) may, as discussed above, be caused by changes in their behaviour from survey to survey.

The depth distribution of the big-eye grunt was similar to that found in the region Benguela-Luanda. This species was common in hauls down to 100 m depth, but the highest catch rates were obtained in the 50–100 m depth range. Other grunts, far less abundant than the big-eye, the bastard grunt (Pomadasys incisus), the sompat grunt (P. jubelini) and the rubberlip grunt (Plectorhynchus mediterraneus) were common in catches in the 10–50 m range, but catch rates were low.

Grunts

The small-sized big-eye grunt is by far the most abundant of the demersal fish on the Angolan shelf. It inhabits the inner shelf from Benguela northwards with roughly equal densities in each of the two regions (Table 7.51). The total biomass representing the mean of all the survey estimates for each region is about 70,000 t. The individual survey estimates show, however, a very high variability. The very low estimates in some of the surveys could well be the result of reduced availability to bottom trawl of this species which at times is found in abundance in mid-water. The true biomass of the big-eye grunt could thus well exceed 100,000 t.

Table 7.51 Big-eye grunt: Estimates of biomass by surveys and regions (1,000 t)

Table 7.52 shows the estimated biomass of other grunts, mainly various Pomadasys species. They are also restricted to shallow waters to the north of Benguela, but are far less abundant than the big-eye grunt with an estimated biomass of about 9,000 t.

Table 7.52 Other grunts: Estimates of biomass by surveys and regions (1,000 t)

The sompat grunt (Pomadasys jubelini) and the bastard grunt (P. incisus) were the most common of the other species of this family.

An analysis of the 1986–89 data (Table 7.53) shows that the big-eye grunt was common in shallow waters (10–50 m) and at intermediate depths (50–100 m), but was seldom caught beyond 100 m depth.

Table 7.53 Benguela-Luanda: Incidence and mean catch of big-eye grunt by depth and surveys 1986–89

The other grunts, except the big-eye grunt were common in the shallow water range (10–50 m), were found occasionally in the 50–100 m zone, but never beyond 100 m bottom depth, as can be seen in the incidence rates in Table 7.54.

Table 7.54 Benguela-Luanda: Incidence and mean catch of Pomadasys species by depth and survey

Seabreams

Table 7.55 shows the time series of biomass estimates, but with the assumed biased autumn estimates for the northern regions omitted. The mean biomass density for the small area Cunene-Tombua was higher than in the northern regions, which demonstrates the high productivity also for demersal fish of this region of intensive upwelling. In this area the seabreams consisted mainly of the large-eye dentex, but with some red pandora in shallow waters. From Benguela northwards, the species composition is much more diversified.

A part of the variability of the estimates in the Cunene-Tombua region could be the effect of shifts in the distribution of a common stock of large-eye dentex covering northern Namibia and southern Angola. This stock has been exploited in a trawl fishery with catches reported from ICSEAF Division 1.3 and 1.4 ranging from 3,000 to 12,000 t in the period 1973–86. The reported landings of large-eye dentex in area 47 declined from about 9,000 in 1986 to about 6,000 t in 1984 through 1986, 4,000 t in 1987 through 1989 and nearly 0 t in 1990–91 (FAO, 1993). The increased biomass estimate of seabreams in the southern region in 1992 since early 1989 could be a response to a decline in exploitation.

Table 7.55 Seabreams: Estimates of biomass by surveys and regions (1,000 t)

Cunene-Benguela

The seabreams were dominated by large-eye dentex (93%).

An analysis of the 1986 and 1989 data showed that large-eye dentex was found mainly between 50 and 200 m, but could be traced to 30 m bottom depth (Table 7.56).

Table 7.56 Cunene-Benguela: Incidence and mean catch of large-eye dentex by depth and year

Benguela-Luanda

The seabreams were represented by a number of species. The 1986–89 species composition was 50% red pandora (Pagellus bellottii), 15% large-eye dentex (Dentex macrophthalmus), 15% Angola dentex (D.angolensis) and 7% Barnard dentex (D. barnardi).

Red pandora was very common at intermediate depths during all surveys, as shown from the high incidence rates in Table 7.57. In the shallow waters, it occurred in more than 50% of the hauls, but with low catch rates.

Table 7.57 Benguela-Luanda: Incidence and mean catch of red pandora by depth and surveys

The other Dentex species were common in both the 50–100 and the 100–200 m depth ranges (Table 7.58).

Table 7.58 Benguela-Luanda: Incidence and mean catch of Dentex species by depth and surveys

Luanda-Cabinda

The number of seabream species was higher than in the region south of Luanda. The 1985–89 species composition was on average: Pagellus bellottii 30%, Dentex angolensis 18%, Dentex congoensis 12%, Dentex canariensis, 10%, Dentex gibbosus 7%, Dentex barnardi 3%, Dentex macrophthalmus 2%, Sparus caeruleosticus 7%, Boops boops 4%.

Table 7.59 Luanda-Cabinda: Incidence and mean catch of red pandora by depth and survey

The incidence rates for the 1985–89 surveys show that the red pandora was most common in the depth range 50–100 m where it was found in most of the hauls (Table 7.59).

For the Dentex species the highest mean catch rates were obtained deeper than 50 m where this group was represented in most hauls, as shown by the high incidence rates in Table 7.60.

Table 7.60 Luanda-Cabinda: Incidence and mean catch of Dentex species by depth and surveys

Croakers

Table 7.61 shows a summary of the biomass data for croakers. In the south this group is represented by the African weakfish found down to 200 m depth, while there are a number of species in the northern regions especially north of Luanda where the genus Pseudotolithus is especially important. The time series of the biomass estimates indicate a possible decline of the abundance of this group from 1985/86 to 1989 followed by a recovery in 1991/92. This may have been an effect of changes in the fishery for the species of high commercial value of croakers. The estimated mean biomass over the period of survey is about 20,000 t.

Table 7.61 Croakers: Estimates of biomass by surveys and regions (1,000 t)

Cunene-Benguela

In 1986 the African weakfish was found mainly between 50 and 100 m, but in 1989 it was found at greater depths (Table 7.62).

Table 7.62 Cunene-Benguela: Incidence and mean catch of African weakfish by depth and year

Luanda-Cabinda

Croakers, mainly longneck croaker (Pseudotolithus typus), cassava croaker (P. senegalensis), bobo croaker (P. elongatus), Angola croaker (Miracorvina angolensis) and Canary drum (Umbrina canariensis), were common north of Luanda, especially in the shallow waters where the group was represented in most of the hauls (Table 7.63). Also beyond 50 m the group was relatively common with typical incidence rates around 50% from 50 to 100 m depth. The blackmouth croaker (Pentheroscion mbizi) was common in deeper waters. Mean catch rates, between 35 and 90 kg/h, were relatively consistent in the shallow waters.

Table 7.63 Luanda-Cabinda: Incidence and mean catch of croakers by depth and surveys

Groupers

Groupers are commercially valuable fish, but restricted to the northern part of the shelf and less abundant. As shown in Table 7.64 the total biomass was estimated at only 2,000 to 3,000 t. This is, however, likely to be an underestimate as the groupers often have a preference for a hard, untrawlable bottom.

Table 7.64 Groupers: Mean biomass estimates by regions and survey periods (1,000 t)

Luanda-Cabinda

Groupers were mainly represented by the white grouper (Epinephelus aeneus) and the dusky grouper (E. guaza). Typical incidence rates were in the order of 35–40% in waters shallower than 100 m. Beyond this depth the group was less common.

The Benguela hake

The Benguela hake was found from the Cunene River to Tombua together with Cape hake and as the only hake species on the slope further north from Benguela to Cabinda.

Table 7.65 shows the mean catch rates for the two regions northern and by depth intervals for survey periods. In nearly all cases the highest rates were from the 300–500 m depth range. There is a sharp decline in catch rate below 500 m. In the 400–500 m range there was a tendency for the catch rate to be somewhat higher on the grounds between Cabinda and Luanda than on the southern grounds. This indicates that the hake had its main distribution somewhat deeper in the northern than in the southern Angolan grounds.

Table 7.65 Benguela hake: Mean catch rates by depth ranges and survey period (kg/h)

Table 7.66 shows the biomass estimates by regions and survey periods. There was a decline of about 50% from 1985/86 to 1989 with some increase to 1991, but again a low estimate in 1992. There is no reported directed fishery for Benguela hake, however, it is caught as a by-catch in the deep-water shrimp fishery and it is likely that at least part of the observed decline in the biomass of Benguela hake has been caused by the increasing effort in the shrimp fishery.

Table 7.66 Benguela hake: Biomass estimates by survey-periods (1,000 t)