3. THE FISH RESOURCES, COMPOSITION, DISTRIBUTION AND ABUNDANCE.

3.1 Introduction

Based on their behaviour fish species are generally classified as demersal or pelagic. Although many demersal fish often occur in mid water and pelagic fish near the bottom this is still a useful classification. In presenting the survey results we use for convenience one further classification by subdividing the small schooling pelagic fish into those belonging to the clupeid - and engraulid families: sardines, sardinellas and anchovies denoted pelagic type 1, and horse mackerels, jack mackerels, hairtails, small scombrids and barracudas denoted pelagic fish type 2. The larger species of pelagic fish, tunas and billfishes are for methodical reasons not included in the survey.

The environmental regimes described under Chapter 2 above largely determine the species composition and abundance of the fish communities contained within them. Thus in the area south of Porto Alexandre dominated by the Benguela Current regime the characteristic species are few, but of relatively high abundance. They include the sardine, Sardinops ocellata, the Cape horse mackerel, Trachurus capensis, the Cape hake, Merluccius capensis , and the large eyed dentex, Dentex macropthalmus.

In the main Angola Current regime the number of species both pelagic and demersal are much more numerous. The most abundant in this pelagic community are the two sardinellas, round sardinella Sardinella aurita and flat sardinella, S.maderensis, furthermore the Cunene horse mackerel, Trachurus trecae, various jack mackerels: Decapterus -, Caranx-, and Chloroscombrus species, hairtails Trichiurus sp, spanish mackerels Scomberomorus sp and barracudas Sphyraena sp.

The demersal fish on the Angolan proper shelf are dominated by species of the families POMADASIDAE, especially Brachydeuterus auritus, SPARIDAE and SCIAENIDAE.

In the deep water community along the slope in the regime of oxygen deficiency and low temperature the abundant fish are the deep water hake Merluccius polli and the greeneyes Chloropthalmus sp. found together with various shrimp species : Aristeus viridens, Parapenaeus longirostris and Nematocarsinus africanus. Various species of small mesopelagic fish are common and probably form an important part of the food chain in the waters outside the shelf.

As mentioned under 2 above the Angolan Sea forms part of the south east Atlantic upwelling system dominated by the Benguela Current. Its main features of currents and productivity are similar to and mirrors the system off north west Africa dominated by the Canary current. The Angolan coast from Cunene to Cabinda have certain ecological features which makes it comparable to the coastline from somewhat north of Cape Blanc in Mauritania to Guinea. Many of the fish species are similar and we will later make a comparison of the production of some of the fish stocks in the two areas.

In our description of the findings of the surveys we have divided the coast into regions which relate to the ecological regimes as follows:

The deep water resources on the slope are mainly found in regions 3 and 4.

3.2 Survey results

The acoustic integrators technique provides in the first instance observations on fish density which when plotted in charts illustrate the relative distribution of the categories of fish. Such charts of relative distribution are shown in Figures 4 an 5 for Pelagic Type 1-species and Figures 8 and 9 for the carangids etc. The acoustic data can then be converted to estimates of biomass for which information on fish size and sound reflection properties is used.

In the following, the survey results in terms of distribution, composition and abundance of the resources are described with reference to the distribution charts, the biomass estimates and the results of the fishing experiments.

3.2.1 Small pelagic sardine-like fish, pelagic type 1.

Namibian sardine, (Sardinops ocellata).

This species was found in the region Cunene-Porto Alexandre only. Its main distribution is Namibia, but as mentioned under 2.2 above, that regime extends into Angolan waters with an oceanographic front located somewhere near the Cunene river in February-March (southern summer) and off or beyond Porto Alexandre in August-September (southern winter). Figure 4 shows the distribution of small sardine like fish between Cunene and Benguela in each of the six surveys. Appreciable quantities of sardine was found only in Survey 3 in mid August 1985 when the species occurred in schools and layers over an approximately 20 by 10 nm area in the inner part of the shelf off Bahia dos Tigres.

This sardine area was surveyed twice and the resulting biomass estimates were both about 120.000 tonnes. A number of mid water trawl hauls gave catch rates exceeding 1 tonnes/hr one even reaching 17 tonnes/hr. The size compositions showed modes of 23 and 27cm, probably representing different age groups and the sardine was found to be in pre-spawning and active spawning condition.

Sardine was also caught in this area in Survey 1 in early February, when it was found together with round sardinella in some smaller quantity, and a few appeared close to the Cunene River in Survey 4 in early November 1985.

The sardine is thus a resource with availability in Angolan waters which at least with the present depleted state of the stock is limited to a short season only, in winter-early spring. Converted to mean standing biomass over a yearly period the survey data indicate a quantity of about 30.000 tonnes.

The occurrence of sardine in Angolan waters both in terms of length of season and total biomass is no doubt largely influenced by the state of abundance of the total stock. This will be commented on further under 4 below.

The sardinellas.

The two species of sardinella, Sardinella aurita - round sardinella, and Sardinella maderensis - flat sardinella were always identified and estimated separately in the catches in which they occured. Separate assessments of their biomass can ,however, only be roughly indicated since the sampling in the areas where their distribution overlapped is not adequate for precise separate estimations.

In our surveys round sardinella was found from Region 1, Cunene - Porto Alexandre northwards to Cabinda. Flat sardinella was caught in only small quantities south of Benguela. The distribution of both species extend northwards into the waters of Congo and Gabon.

Figures 4 and 5 show the distribution of pelagic type 1 fish from all six surveys. These maps will be commented on jointly with Table 3 which presents the biomass estimates by surveys and regions.

In Regions 1 and 2 appreciable amounts of pelagic 1 fish were only found in the first three surveys. As referred to above the good concentration of pelagics found off Bahia dos Tigres in Survey 3 consisted of sardine. Round sardinella was however probably the main part of the fish found in about the same locality during Survey 1, and also the scattered and small areas of pelagic 1 fish located along the coast up to Benguela in the other surveys consisted mainly of sardinella. In Survey 2 very high densities were observed in some of these school locations in Regions 1 and 2 thus accounting for the relatively high biomass estimates, the reliability of which are however low. In general as demonstrated both by Table 3 and Figure 4 the abundance of sardinellas in the two southernmost regions must be termed sporadic and low.

The main biomass of the Angolan sardinellas was in all surveys found in Region 3, Benguela - Luanda and Region 4, Luanda - Cabinda. This is demonstrated by the distribution of pelagic 1 fish for these regions, Figure 5. Apart from some minor contributions to the total biomass from anchovy and west African ilisha inshore mainly in Region 3, these charts represent the two sardinella species.

An important feature of these charts is that the character of the fish distributions changes through the programme. In Surveys 1 and 2 the sardinellas are found in well defined school areas with dimensions alongshore of 30– 40 nm or more and reaching 20 - 30 nm offshore. From Survey 3 on, the areas of distribution are smaller and dispersed especially for the denser concentrations. This change is probably related to the decline in total abundance observed over this period, and the change in the composition of the stocks with a reduced proportion of young and juvenile fish.

When deducting the biomass of the Namibian sardine from the total pelagic fish 1 estimates in Table 3 the remaining part which is predominantly sardinella declines from about 500 thousand tons in the first two surveys to between 260 and 350 tons over the remaining surveys. There is a need to discuss whether this represent a true change. The most important problem in acoustic assessments of sardinellas is the behaviour of surface schooling and vessel avoidance. Such behaviour was observed particularly in Surveys 3 and 6. The schools were recorded by sonar, identified and converted to standardized units. The total numbers for Survey 3, 5 and 6 were 860, 600 and 940 respectively. With about the same biomass estimate for the three surveys, Surveys 3 and 6 thus had about 50% more schools recorded than Survey 5. This supports the idea that the degree to which surface schooling occurs may vary and cause variations in the assessments. Persistent surface schooling has also been observed by Marchal of flat sardinella over the outer parts of the shelf off the Congo.

The sardinella stocks are known to migrate north in the southern winter period. During 1985 the shelves off the Congo and Gabon were included in the four surveys. Total biomass estimates of all small pelagic fish (types 1 and 2 ) for these four surveys were 100, 90, 200, 55 thousand tons respectively. Sardinellas were, however, only found in appreciable quantities in Survey 3 in September and were mostly juvenile fish less than abt 20 cm total length. A northward shift during August-September seems also to be demonstrated in Table 3 and may explain the low biomass estimate for Survey 3.

The estimates for Surveys 5 and 6 should however seasonally be comparable to those of Surveys 1 and 2. They show a decline of biomass of nearly 50 percent. Although the true decline is probably somewhat less since the last surveys were more troubled by surface schools than the two first, and accordingly affected more by underestimation, one must conclude that a significant decline took place in the biomass of the sardinellas from January- June 1985 to January - June 1986 from a level of about 500 000 tons to about 300 000 tons. The mean standinng biomass during 1985 was about 400 000 tons.

The catch records provide some information on the relative abundance of the two species. In the total of 296 mid water trawl hauls round sardinella was caught in 115 hauls and flat sardinella in 145 hauls with mean catch rates of 97 - and 60 kgs/hr respectively. This gives a ratio round to flat sardinella of 0.8 as regards incidence of occurence and 1.6 in catch rates. The ratios of proportion in the total catches is 1.33. There is some variation in these ratios between surveys but little variation between Regions 3 and 4 as shown in Table 4. For Surveys 4 and 5 the proportion of round sardinella seems to have been somewhat higher than in the others. One may, however question whether these ratios in the catches reflect the true proportions of the species. The catchability in mid water trawl is very low for both round and flat sardinella and even at night they avoid the gear or escape through the meshes to a large extent. Small differences in catchability or selectivity would affect the ratios. But on the evidence of the ratios one would conclude that there is an approximately equal biomass of each species.

There is as expected some evidence of a difference in depth distribution. Both species are caught out beyond 100 m of depth, but with only small catches of flat sardinella outside of 50 m depth.

Table 4. Ratios of round to flat sardinella in the catches.
i = ratios of incidence in mid water hauls
w = ratios of mean catch rates

Figure 6 shows the pooled size distributions of round sardinella from Regions 3 and 4 for each of the surveys. Medium sized and juvenile fish were present during almost all of the first year of surveys, but the round sardinella of Surveys 5 and 6 consisted only of adult fish of abt. 25 to 35 cm of length. For the flat sardinella, see Figure 7 there is also less medium sized and juvenile fish in Surveys 5 and 6 than in Surveys 1 and 2. This could be an indication of a reduced recruitment for both species in 1986 as compared with 1985. The round- and flat sardinella found off Congo - Gabon in Surveys 3 and 4 were all juveniles below 20 cm of length. This is likely to be an area of recruitment for the stocks, but our data does not allow an assessment of its relative importance.

3.2.2 Other small pelagic fish. ( Pelagic type 2 ).

The distribution charts for this category are shown in Figures 8 and 9 and the estimates of total biomass are recorded in Table 5. A number of species are included within this group and estimates of their contribution to the total biomass can only be roghly indicated.

The horse mackerels represent the largest component of the category with two species, the Cape horse mackerel Trachurus capensis, and the Cunene horse mackerel Trachurus trecae. The Cape horse mackerel was found almost exclusively to the south of Porto Alexandre while the Cunene horse mackerel's distribution cover the whole of the Angolan coast and range into the Congo and Gabon.

Cape horse mackerel.

The two horse mackerels account for nearly the total of the pelagic fish 2 biomass in Region 1. Their proportion in the catches of the mid water - and bottom trawls are set out in Table 6 for each survey. The two species are quite similar and their may have been some problems of identification during Survey 1. But this cannot explain the trend shown in the table of a shift in the proportions in favour of the Cunene horse mackerel. For the assessments we will assume that the ratios of Cape to Cunene horse mackerel in Region 1 was 9 : 1 during 1985 and 2 : 8 in Surveys 5 and 6. The biomass estimates of Cape horse mackerel in Angolan waters over the six surveys are thus as follows: 110, 90, 200, 290, 20, and 10 thousand tons. It is probable that the variations in the biomass estimates between surveys in part reflect a seasonal fluctuation in the availability of this resource in Angolan waters as for the sardine. In winter and spring the available standing biomass was 200 - 300 000 tons. The mean standing biomass for 1985 was about 170 000 tons. When comparing the estimates for Surveys 5 and 6 with those of Surveys 1 and 2 in the same season of the year, there is evidence of a considerable decline. Since southern Angola represents the northern boundary of distribution of this species, especially in summer, this decline could well be caused by a minor shift in the migration pattern of the stock.

As shown in Figure 8 the distributions over the relatively broad shelf off Baia dos Tigres had their highest concentration over the outer parts of the bank and in the four first surveys they seemed to continue south into Namibia. In Survey 4 the highest concentration of horse mackerel was found 2 – 15 n m outside the shelf edge at 100 – 150 m depth over deep water.

Figure 10 shows the size compositions of Cape horse mackerel pooled for each survey. The larger biomasses of Surveys 3 and 4 consisted mostly of small sized fish of less than 20 cm total length. The largest size group caught is about 35 cm.

The Cape horse mackerel is a stock shared with Namibia as is the case for all the main stocks in Region 1. Its partial availability in Angola will be related to its state of exploitation. This will be discussed further in Section 4 of the report.

The Cunene horse mackerel and associated species.

For Region 2, Porto Alexandre to Benguela identification by fishing indicated that nearly all of the “acoustic” biomass should be allocated to the Cunene horse mackerel. From Benguela northwards, however, most of the species which we have included in this group are represented in the catches. But their proportion does not provide a ready basis for an estimation of the biomass of each of them. This is because fishing for identification is not distributed randomly. When a certain type of schools or other formations of fish in an area had been identified and sampled no further fishing would be made in that area. A second important reason is that the species grouped together in this category have different characteristics of distribution and behaviour which affect their catchability for the mid water gear. The West African bumper, Chloroscombrus crysurus thus occur in small relatively loose schools which are fairly easy to catch. The lookdown, Selene dorsalis and the hairtail, Trichiurus lepturus has a similar high catchability and this is also our experience with the barracudas. Small sized horse mackerel also belongs in this category, but this is not the case with the dense schools of adult horse mackerel, false scad, Decapterus rhoncus and chub mackerel, Scomber japonicus which in Region 3 in some of the surveys were found in areas of concentration over the middle part of the shelf. These are like the adult sardinellas almost impossible to catch by day in the high temperatures of the surface waters. When fished by night sufficient numbers are usually caught to allow identification and sampling. The procedure that has therefor been followed for biomass estimation of the main pelagic 2 species has been to allocate the areas of high concentration offshore, see Figure 9 to adult horse mackerel. This will then include minor quantities of chub mackerel and false scad. The remaining part can then be allocated in accordance with the proportions in the catches. It follows from this that the pooled size distributions used in this report are not generally representative of the populations of these species and often only indicate the occurence of the various size components.

The separate estimations of the dense school areas in Regions 3 showed that they accounted for 48, 22, 0, 20, 65, and 0 per cent of the total for Surveys 1 through 6 respectively. Table 7 shows the proportions in the catches of the most common pelagic 2 types of fish for all surveys combined. The Cunene horse mackerel occurs in the largest number of the catches. Hairtail is almost as common, but with smaller proportion by weight. Bumper, barracudas and lookdown are also widespread. Cunene horse mackerel and hairtail show a smaller proportion of the catches in Region 4 than in Region 3 while the opposit is the case for bumper and lookdown.

Table 8 shows the final results of the attempts to allocate the biomass estimates among the various species in terms of mean standing biomass over the period covered by the six surveys.

The Cunene horse mackerel is the dominating species with a mean standing biomass of nearly 200 000 tons. Region 3, Benguela to Luanda represents its main area of distribution, but with considerable variations in its abundance between the surveys. This is mainly caused by the location of dense school areas in some surveys, often over the narrow shelf off Pt das Palmeirinhas as shown in the distribution charts. There is however no evidence of any trend of decline of this stock over the period of the surveys although the estimate for survey 6 was low. The standing biomass for the four surveys during 1985 is thus the same as that for all six surveys.

The occurence of the other species is largely restricted to Regions 3 and 4. Together their standing biomass amounts to nearly 90 000 tonnes.

Figure 11 shows the size distributions of the samples of Cunene horse mackerel for each survey from all regions combined. Juvenile, medium sized and adult fish is present in all regions, and there is no clear seasonal trend, but there was a predominance of juvenile fish in Surveys 5 and 6.

Figure 12 shows the size distributions for samples from Region 3 and 4 for bumper, lookdown, barracudas and false scad.

3.2.3 Demersal shelf resources

The acoustic method is less favourable to investigations on demersal biomass, as the fish close to the bottom tends to escape detection by the echosounder. In addition, demersal resources are often distributed in a more scattered pattern as compared to the pelagic resources and this creates problems in relation with the resolution of the system. For the above reasons, the acoustic method often underestimates the demersal resources.

The demersal fish can be better assessed with trawl surveys. A method commonly used is the swept-area method, where randomly distributed and standardized trawl hauls on the shelf are used as samples to obtain a mean fish density in the area investigated. This density is later converted to biomass by multiplying it with the area of the shelf concerned.

In the course of the Fridtjof Nansen program, swept-area investigations were carried out during all surveys in region 4, in the last three surveys in region 3, only in the last two surveys in regions 1 and 2. The so-called semi-random sampling program was adopted, i.e. the trawl stations were randomly set out at daytime along the given acoustical survey track.

Biomass estimates for regions and surveys where random hauls are missing, were obtained by using the bottom trawl hauls carried out in connection with the acoustic investigations with the exclusion of all pelagic species in the catches from the calculations. These bottom hauls, used for identification purposes, tend to give overestimates as they often are aimed trawling on acoustic registrations. However, if the pelagic species are excluded, this effect is much less pronounced and comparisons have shown that the estimates from these hauls are reasonable, but should perhaps be interpreted as optimistic.

The following table gives biomass estimates by regions and surveys (thousand tonnes)

* = Calculated from bottom hauls carried out in connection with the acoustic survey.

In Annex Report, tables of the processed data from the trawl surveys are given by species, surveys, regions and bottom depth zones. Below, the main conclusions, based on this material, are given by main geographical area.

River Cunene to Porto Alexandre.

The main demersal species in this region on an annual basis is Dentex macrophthalmus (the large-eye dentex), representing 33% of the total demersal catch togwther with Merluccius capensis ( the Cape hake) 9%.

No drastic seasonal changes in the demersal fauna were observed, and the two above mentioned species dominated the fauna during all surveys, except that the hakes importance declined during the two last coverages. The main species and their relative importance expressed as percent of total demersal catch are:

The highest densities of the hake were found beyond the 100m depth contour, while big-eye dentex were found mainly between 50 and 200 m bottom depth.

Combining the annual mean biomass estimate for the region with the annual species composition a rough estimate of 12 thousand tonnes of large eye seabream, and 3.5 thousand tonnes of hake are obtained.

Below, some biomass estimates, based on all bottom trawl stations in the area are given for the most important species (in thousand tonnes):

As these estimates are based on all bottom trawl stations they should be interpreted as somewhat optimistic.

Porto Alexandre to Benguela

This region has a very narrow shelf and only a total of 18 bottom trawl stations were carried out, which is not sufficient to give species distribution in the catches by surveys. Based on all 18 hauls, the species distribution is given below (in percentage of the total catch:

With a mean annual biomass of 24 thousand tonnes, the area should tentatively hold 9 thousand tonnes of Dentex macrophthalmus. The Umbrina, Pteroscion and Pagellus add up to 3.6 thousand tonnes.

From Benguela to Luanda.

A drastic change in species composition was found in the demersal communities of this region as compared with the southern regions. While Dentex macrophthalmus has lost its dominance, and Merluccius capensis is totally missing from the catches, Brachydeuterus auritus has become the dominating species, representing 39% of the total demersal catch, followed by Pagellus bellottii (8%) Dentex macrophthalmus (6%) and Umbrina canariensis (4%).

The main species composition in the region by surveys, expressed as percent of the total demersal catch is given below:

The table above indicates a seasonal migration of Dentex macrophthalmus, a species with a modest presence on an annual basis, but important in the catches of surveys 3 and 4. Whether this species migrates in from the south or from depths below 200 m, cannot be shown from our investigations.

Brachydeuterus auritus is mainly caught between 30 and 100m bottom depth, together with Pagellus bellottii and Pomadasys incisus. The shallower depths less than 30 m are mainly inhabited by Brachydeuterus auritus together with Galeoides decadactylys and Pseudotholithus typus. The depths between 100 and 180 m are preferred by Dentex angolensis and Umbrina canariensis.

Estimates of biomass by species and surveys are as follows (thousand tonnes):

The yearly mean estimates by species groups are (in thousand tonnes):

Luanda to Cabinda

This region is where the demersal biomass was most intensively surveyed. Swept area investigations were carried out during all coverages and a total of 237 random trawl stations were worked out.

The main species on a yearly basis is Brachydeuterus auritus, 37% of the total catch, followed by Pagellus bellottii (7%) and Dentex angolensis (6%). The seabream group as a whole makes up 23%of the total catch of demersal species, the croakers 10% and the bigger grunts, excluding the Brachydeuterus auritus, only 4%. This area thus holds a farly high percentage of good quality fish and it is probably the most interesting concerning exploitation by bottom trawl.

The main species distribution during the various surveys is (in % of the total catch during the survey):

On an annual basis the seabreams seem to inhabit mainly the 30 to 200m bottom depth zone, while the croakers have their highest density in depths shallower than 50 m.

Estimates of biomass by species groups and surveys are as follows (thousand tonnes):

3.2.4 Deep water slope resources

The slope between 200 and 600m bottom depth is about 4750nm², and holds valuable resources of hake and shrimps. The distribution of the area of the slope by region is given below(nm²):

In the two southern regions, most of the slope is too rough for trawling. There are several areas suitable for trawling in area 3 and area 4 holds the most extensive grounds for deep sea fishing.

Fishing on the slope was carried out only from the third coverage of the program. At the beginning, shrimps were not sorted and identified by species, which was done in full during the three last coverages.

The work on the slope is distributed as follows(number of trawl stations):

There does not seem to be any major difference in the deep water fish fauna of the slope in the northern area as compared to the southern area, neither have we been able to find seasonal fluctuations.

The deep water fauna consist of Merluccius polli, Myctophidae, Synagrops microlepis, Chlorophthalmidae, sharks, shrimps, squid, and various deep water fishes without commercial value.

Three species of shrimp are most prominent, the striped red shrimp (Aristeus varidens), deepwater rose shrimp (Parapenaeus longirostris) and the spider shrimp (Nematocarcinus africanus), the last of no commercial importance.

The proportion by weight of the species in the deep-sea catches is as follows (Myctophidae and Synagrops microlepis excluded):

The mean densities in regions 3 and 4 are 30.2 and 19.5 t/nm² respectively and the corresponding total biomass estimates are 52 and 41 thousand tonnes.

The estimated mean density of biomass between regions 3 and 4 differs considerably, as region 3 has 2.2 and 1.6 times higher density in the bottom ranges 200–300 and 300–400 m respectively, while in deeper waters the estimates coincide. It would be interesting to correlate these observations with the fishing effort in the two regions, but we have no such information available.

The biomass distribution in regions 3 and 4 is (thousand tonnes):

To see to what extent the different groups coccur in the catches it is useful to look at the mean density by bottom depth strata in region 3 and 4 pooled together (tonnes/nm²):

The highest densities of shrimp were located in the 400–500 m zone. In this area the mean catch of Merluccius was 24 times the catch of the shrimp. The hake catches were never less than 4 times the catches of commercial shrimps in any zone.

Biomass estimates by depth zones in region 3 (tonnes ) :

Biomass estimates by depth zones in region 4 (tonnes):

The above tables indicate that 83% of the total biomass of hake is located between 300 and 500 m in both regions.

Length frequency distributions from all samples of Benguela hake are shown in Fig. 13. The size range is from 14 to 45 cm, with a clear mode around 29 cm.