5 SURVEYS IN THE SOUTHWEST INDIAN OCEAN

In the 1970s and early 1980s, the development of marine fisheries in countries bordering the Southwest Indian was of great interest to donor agencies, such as NORAD and UNDP. After the termination of the strong co-operation with FAO and UNDP in the Arabian Sea, NORAD decided to deploy the DR. FRIDTJOF NANSEN in Mozambique, where it supported a number of long-term fisheries development programmes.

After the initial surveys in Mozambique, an interest was developed to also cover the adjacent states, usually in co-operation with FAO and local FAO/UNDP or NORAD projects.

A total of 16 surveys were conducted in the period 1977–90, of which seven in Mozambique, with 13 complete or partial coverages of the shelf areas, four in Kenya, three in Tanzania and one each in Madagascar and the Seychelles.

The survey in the Seychelles was incomplete, while the one off Madagascar was mainly for oceanographic purposes, covering only the southern part of the island. The results of the surveys in Kenya, Tanzania and Mozambique are described and discussed below.

5.1 KENYA, 1980–83

Survey objectives and effort

Surveys of the shelf of Kenya was part of the DR. FRIDTJOF NANSEN's East African Coast programme in the early 1980s to investigate small pelagic fish with acoustic methods and demersal fish with bottom trawling. The four surveys in Kenya, in December 1980, August and December 1982 and May 1983 covered together all trawlable parts of the shelf and the slope from about 10 m to 500 m. The shallow, more productive part of the shelf was covered in each of the surveys. The results were briefly described in cruise reports (IMR, 1982d; Nakken, 1981; Iversen, 1983) and summarized in a special report for the “NORAD-Kenya Seminar on the Marine Fish Stocks and Fisheries in Kenya” held in Mombasa in 1984 (Iversen, 1984; Iversen and Myklevoll, 1984b).

Table 5.1 shows the operational data of the four surveys. The degree of coverage for the acoustic investigations, was generally high and particularly so for the August 1982 survey. The trawl stations are those recorded as successful swept-area hauls, available in the NANSIS data bank with the exception of those from the 1980 survey.

Table 5.1 Details of the surveys in Kenya

Figure 5.1 shows the shelf of Kenya, the sub-areas used in the trawl survey programme and the coverage in the August 1982 survey. Table 5.2 shows the areas of the depth strata by subarea (Iversen, 1984). Most of the southern area is very deep and this part was only covered in the December 1980 survey. The North Kenya Bank is narrow with a steep slope. The Malindi Bank-Ungama Bay area has the widest shelf with a generally smooth trawlable bottom. The bottom trawl investigations in the 1982/83 surveys were mostly confined to this area and to the southern part of the North Kenya Bank.

Figure 5.1 The shelf of Kenya: Course tracks in the August 1982 survey and the investigated areas. 1: North Kenya Bank; 2: Malindi-Ungama Bay area; 3: The southern area

Table 5.2 The shelf of Kenya by sub-areas and depth strata (nmi²)

The species diversity in this region is high and the task of identification was aided by the participation of Mr. S.C. Venema from FAO and Dr. P.C. Heemstra from the J.L.B. Smith Institute of Ichthyology in South Africa in the December 1980 survey during which some 260 species of special taxonomic interest were collected (Nakken, 1981). During this cruise the need for inputs of taxonomists was realized and it can be considered as the start of a close co-operation between the DR. FRIDTJOF NANSEN programme and FAO's Fish Identification Programme (Venema, 1981).

Hydrography

Hydrographical profiles worked off the southern, central and northern coasts showed no drastic changes in the seasonal environment over shelf. The temperature of the surface layer increased from 25°C in August 1982 to 27–28°C in December 1982 and was again about 28°C in May 1983. The thermocline was found at 100–150 m in August 1982, at 50–125 m in December 1982 and at 75–100 m in May 1983. The surface layer had high oxygen content, more than 4 ml/l, in all seasons.

Pelagic fish

Charts of the distribution of echo integrator outputs corresponding to density of fish recorded in mid-water (Iversen, 1984) showed that pelagic fish was present over wide parts of the shelf, but mostly scattered. The aggregations that were found were not of high density, and were confined to the central Malindi-Ungama Bay area in all surveys.

The biomass of the fish observed in mid-water was estimated with the acoustic integration method with the following results (Iversen, 1984):

These estimates include both small pelagics and semi-demersals such as ponyfish, but they do not represent the total biomass of these types of fish since the surveys did not include the shelf area from 0 to 20 m depth which is often a productive zone. The mean density in the shelf area observed between 20 and 200 m was, however, low with an average of about 15 t/nmi². No areas of high aggregations were found and the reports only mentioned occasional observations of surface schools.

Table 5.3 Kenya: Catch rates and incidence of pelagic fish in 42 hauls by bottom trawl between 10 and 50 m depth in the three 1982–83 surveys

Pelagic fish also formed part of the catches with the bottom trawl in shallow waters. As shown in Table 5.3 these data indicated that Clupeidae and Carangidae were the most common taxa, followed by Scombridae and Sphyraenidae.

There was a high species diversity. The most common pelagic species were:

Demersal fish

The mean catch rates of demersal fish by depth ranges are shown in Table 5.4. The deep water areas (200–500 m) were covered in December 1980 and August 1982. The catches consisted mostly of various deep-water fish Zeniontidae, Chlorophthalmus sp. and Diaphus sp., sharks and squids, most of which would seem to be of little commercial interest. Small catches of deep-water lobster species (Palinuridae) were taken at 280–350 m depth. The catch rates varied from one to a few kilogrammes per hour in ten hauls and with one catch of 25 kg/h. More detailed information on deep-water crustacean resources off Kenya is available from other surveys as reported by Mutagyera (1984).

In the two last surveys the trawling was mainly restricted to the shelf inside 200 m depth off Malindi and the Ungama Bay. The most common families among the demersal fish were snappers, grunts, groupers and mullets, which have been included in the category demersal fish in Table 5.5. The semi-demersal ponyfishes were particularly abundant, while one-third of the catch consisted of pelagic fish with a species composition as described in Table 5.3.

Table 5.6 shows the biomass estimates by sub-area for each survey as summarized by Iversen (1984). The high estimate for the North Kenya Bank in the last survey is due to a single catch, blue-spotted jobfish (Pristipomoides filamentosus) which appears to occur in shoals in deep water. Otherwise, the highest biomass is found in the central Malindi-Ungama Bay area. Demersal fish of commercial interest formed, however, a small part of the catches in this area as shown in Table 5.5. An estimate of the composition of the biomass in this area in the 10– 200 m depth range based on the three 1982/83 surveys was as follows:

Table 5.4 Kenya shelf: Total catch rates of bottom trawl hauls by depth strata (kg/h)

Table 5.5 Sub-area Malindi-Ungama Bay: Mean catch rates of bottom trawl hauls by survey of main species groups on the shelf 10–200 m (kg/h)

Table 5.6 Kenya: Swept-area biomass estimates by survey and sub-areas

Summary of findings

The survey covered the shelf outside the inshore reef zone from a depth of 10–15 m to about 500 m. Most of the trawl hauls were in the 20–200 m depth range. The shallow grounds inside the reefs which in the central part are extensive and cover about 250 nmi² could not be covered by the vessel. The surveys did not include large pelagic fish, tunas and tuna like fish in offshore waters.

The estimates of fish biomass with the acoustic method showed a range of 18,000 to 32,000 t with a mean of 25,000 t. Samples showed a large variety of species of various typical pelagic families, but the estimates also included semi-pelagic ponyfish. Most of the fish was very scattered and located inside 200 m depth in the central Malindi-Ungama Bay area.

The swept-area estimates showed a range of 14,000–16,000 t total biomass. On the central shelf in the 15–200 m depth range about 70% of the bottom trawl catches consisted of pelagic fish and ponyfish and only about 12% were reas demersal species such as snappers, grunts, groupers and mullets.

The two methods of biomass estimation cover partly the same groups of fish. This is especially so for the pelagic families and ponyfish. A combined estimate may indicate a total standing biomass of about 35,000 t, of which by far the greatest part would be small pelagics and ponyfish. This represents a mean density over the shelf of about 20 t/nmi², a level consistent with the low-productive tropical environment. However, the inshore reef areas, which form the main grounds of the artisanal fisheries are likely to have a higher productivity. On the basis of the DR. FRIDTJOF NANSEN surveys, however, participants at the 1984 seminar drew the conclusion that the resources which had been identified and estimated did not warrant the development of industrial fisheries. Kenya's landings of marine fish have remained stable in the period 1982–92 at about 7,000 t.

5.2 TANZANIA, 1982–83

Survey objectives and effort

The Tanzanian shelf formed part of DR. FRIDTJOF NANSEN's East African Coast programme in the early 1980s to investigate small pelagic fish with acoustic methods and demersal fish with bottom trawling. In each of the three surveys, June-July 1982, November-December 1982 and May 1983 the trawlable parts of the shelf from the Zanzibar Channel south to the Rufiji delta (sub-areas 2 and 3) were covered by a bottom trawling programme. The whole shelf was also covered three times by acoustic surveys, except for the narrow shelf south of Kilwa Kivinje in the third survey. The surveys and their main findings were briefly described in cruise reports (Myklevoll, 1982b; IMR, 1982c and 1983a) and summarized in a special report for the “NORAD-Tanzania Seminar to Review the Marine Fish Stocks and Fisheries in Tanzania” held in Mbegani, Tanzania in March 1984 (Iversen and Myklevoll, 1984a and Iversen et al., 1984).

Table 5.7 shows the operational data for the three surveys. The degree of coverage for the acoustic investigations was high in all surveys.

Figure 5.2 shows the Tanzanian shelf and the sub-areas used to analyse the trawl data. Table 5.8 shows the extent of the depth strata. The sub-areas 2 and 3 identified as “Zanzibar” and “Mafia” together comprise nearly 90% of the shelf inside 200 m depth and include the main trawlable parts, therefore the bottom trawl surveys were mainly confined to these sub-areas.

Like Kenya, Tanzania has also a high diversity of marine species and the taxonomic work was supported by the participation in the second survey of Ms Gabriella Bianchi of FAO's Department of Fisheries. This survey also provided a first opportunity to test drafts of the Species Identification Sheets for the Western Indian Ocean, produced by FAO (Fischer and Bianchi, 1984).

Figure 5.2 The investigated sub-areas of the Tanzanian shelf: 1) Pemba, 2) Zanzibar, 3) Mafia and 4) the southern area and survey routes of the November-December 1982 survey

Table 5.7 Details of the surveys in Tanzania

Table 5.8 Estimates of the areas of the shelf of Tanzania by depth strata (nmi²)

Hydrography

Sets of hydrographic profiles were worked from the coast in all surveys; the full data are presented and discussed in Iversen et al. (1984). The depth of the upper-mixed layer shows little change between surveys. The oxygen content of the shelf and slope waters was relatively high and is unlikely to have affected fish distribution. Surface salinity was low in May following high river runoffs at that season. The low-salinity water was restricted to inshore areas, an effect of the onshore surface transport set up by the southwest monsoon.

Pelagic fish

Charts of mean echo integrator outputs of the densities of fish in mid-water showed the highest values in localities close inshore and in the channels between the islands and within about 30 nmi of the shore in the central Zanzibar and Mafia areas. The densities of the aggregations were, however, overall low (IMR, 1982c and 1983a and Myklevoll, 1982b).

Acoustic estimates of the biomass of the fish observed in mid-water over the shelf north of 9°S (sub-areas 1, 2 and 3) were as follows: (data from Iversen et al., 1984 but adjusted to a target strength of -34 dB/kg for 17 cm fish)

Semi-pelagics such as ponyfish were included in these estimates and may have represented a considerable part of the biomass.

Hauls with the pelagic trawl gave insignificant catches and did not represent a successful sampling method. Pelagic fish formed, however, part of the catches with the bottom trawl in shallow waters. As shown in Table 5.9 Carangidae and Clupeidae were the most common families, while Scombridae and Sphyraenidae appeared with high incidence, but with low catch rates.

Table 5.9 Tanzania: Catch rates and incidence of pelagic fish in 123 bottom trawl hauls, 10–50 m depth. Mean of all three surveys

The following species were among the most common of the pelagic families:

Demersal fish

Table 5.10 shows the mean catch rates in the bottom trawl for two sub-areas. The catches in the deep range (200–500 m) were generally low and consisted of various lizardfishes and deep-water fishes such as Myctophidae, Chlorophthalmus spp. and Cubiceps spp., apparently of little commercial interest. In a relatively narrow sector at 320–420 m on the slope off Dares-Salaam and a bit further south catches of deep-water shrimp and lobsters were obtained. The mean catch rates were low, but a few hauls gave rates up to 130 kg/h for deep-water shrimp and up to 15–16 kg/h for each of Nephropidae and Palinuridae.

The biomass in the depth range 10 to 200 m as estimated by the swept-area method was as follows, (t):

Table 5.10 Tanzania: Mean catch rates of bottom trawl by depth range, sub-area and survey (kg/h)

On the shelf inside 200 m the most important demersal families of commercial interest were in order of abundance:

The most common demersal species in the catches are listed below by species groups families:

These species are included in the group “demersal fish” in Table 5.11 which summarizes catch rates by main groups in all hauls made on the shelf.

The main species in the bottom trawl catches of the three DR. FRIDTJOF NANSEN surveys in Tanzania have been analyzed on by Yonazi (1985).

Table 5.11 Tanzania: Mean bottom trawl catch rates by species groups, sub-area and survey (kg/h)

In the Zanzibar sector ponyfishes and pelagics formed on average 44 % of the catches, and in the Mafia sector as much as 59%.

Summary of assessments

The surveys covered the shelf from a depth of 10–15 m to about 500 m. Most of the survey effort was spent in the 15–200 m depth range. The shallow grounds inside the reefs which are extensive in the central part of the coast with an area of more than 1,000 nmi² could not be navigated by the vessel. Shallow-water shrimp resources were thus not covered by the surveys nor those of large pelagic fish, tunas and tuna-like fish in offshore waters.

The estimates of fish biomass in mid-water with the acoustic method in the central Zanzibar-and Mafia sectors ranged from 57,000 t to 101,000 t with a mean of 75,000 t. This represented pelagic fish, but included some semi-pelagics such as ponyfish.

The swept-area estimates based on bottom trawling ranged from 45,000 t to 52,000 t in the depth range 15–200 m in the sub-areas Zanzibar and Mafia, with a mean of 47,400 t. Taking into account that 30% of the pelagic fish and ponyfish biomass estimated by the swept-area method is also included in the acoustic estimate, then a combined estimate of the standing biomass is about 110,000 t for the shelf outside the reefs, and about 150,000 t for the whole central shelf if the productivity of the areas inside the reefs, not covered by the surveys, is assumed to be equal to that outside. These estimates correspond to a biomass density of 28 t/nmi², a somewhat higher level than that found for Kenya, but still demonstrating a modest productivity.

Iversen et al. (1984) concluded that the yield of Tanzania's marine fisheries of about 40,000 t might be increased if the fishing areas were extended beyond the reefs, but the prospects for this were perhaps not too good because of rather low fish densities and a predominance of ponyfish. The highest catch rates of both pelagic and demersal fish were, however, usually obtained at depths shallower than 50 m and most of this fish would probably already be available to the inshore fleet. Tanzania's marine landings are reported to have grown from about 27,000 t in 1982 to 55,000 t in 1991 (FAO, 1985 and 1993) through increases of catches of small pelagics, sardinellas and Indian mackerel, but also of demersals. These landings form a high percentage of the biomass as estimated by the DR. FRIDTJOF NANSEN.

5.3 MOZAMBIQUE, 1977–90

Survey objectives and effort

As a part of NORAD's long-term support of fisheries research and development in Mozambique, seven survey assignments with the DR. FRIDTJOF NANSEN were mounted in that country over a period of about 14 years. The particulars of these surveys are set out in Table 5.12.

Table 5.12 Data on surveys in Mozambique, 1977–90

The seven surveys taken together represent a full year of field operation of the vessel, a very considerable effort which delivered a great amount of information and data. This review deals only with the general findings concerning the distribution, composition and abundance of two groups of resources - small pelagic fish and demersal fish - and should be considered as a supplement to the information from many other research activities undertaken by IIP.

Hydrographical investigations, often extensive both in scope and coverage formed part of most of the surveys. However, the results of these investigations will not be included in this review as they are reported in detail in some of the survey reports as well as in special reports by oceanographers of the Instituto de Investigação Pesqueira (IIP), Maputo.

The shallow-water and deep-water shrimp surveys with the DR. FRIDTJOF NANSEN formed part of IIP's ongoing programme of investigations and assessment of shrimp resources. The results are well documented and have been discussed at special shrimp seminars organized in Maputo by the State Secretariat of Fisheries of Mozambique, therefore the findings of these surveys will not be reviewed here. However, in some cases these surveys also included observations of pelagic and demersal fish which may be of interest for this review.

The two research institutions involved in the surveys, IIP and IMR cooperated in survey planning and execution as well as in reporting. The seven surveys were described in cruise reports and analyzed in summary reports, as follows:

See Appendices I and II for full information.

Acoustics

In the review of the state of the acoustic instruments in the various assignments of the DR. FRIDTJOF NANSEN (see Section 2.3) it was concluded that the monitoring of the instruments during Survey I had been incomplete. No instrument reports were available for this period, but it was documented that the source level of the main echosounder had varied between the cruises (IMR, 1978c). The acoustic estimates from survey I are therefore considered less reliable than those from later surveys, therefore they have not been included in this review. The remaining acoustic surveys only cover three quarters of the year, from April to November.

Bottom conditions

The shelf of Mozambique is narrow (10–15 nmi), except in the central part, where it forms the wide Sofala Bank (Fig. 5.3). The much smaller Boa Paz Bank is situated in the northern part of the Delagoa Bay.

Figure 5.3 The shelf of Mozambique and its bottom conditions: 1) impossible to use bottom trawl; 2) possible with caution; 3) good trawl bottom. From Sætre and de Paula e Silva (1979)

The bottom conditions are important for the swept-area trawl method. Figure 5.3 shows that there are large areas of the shelf where bottom trawls cannot be used or must be used with caution. This created problems for representative sampling, and since many types of bottom fishes seem to aggregate on rough bottom, it is doubtful whether trawl catches from trawlable smooth bottom nearby will provide unbiased observations. The effect in a swept-area trawl assessment would be an underestimate of these types of fish.

Table 5.13 shows the area of the shelf by sectors and depth ranges (based on calculations from British Admiralty charts by Sætre and de Paula e Silva, 1979). Especially the Sofala Bank is shallow with about 80% of the area with less than 50 m depth.

Table 5.13 Mozambique: Shelf area by sectors and depth ranges. Recalculated from Sætre and de Paula e Silva (1979) (nmi²)

The narrow shelf north of Angoche was only covered in survey No. I. Subsequent surveys were limited to the Sofala Bank and included twice the Delagoa Bay. In the last survey the slope of the shelf from 17°S southwards was covered to assess deep-water shrimp.

Species identification was, for the early surveys, mostly based on Smith (1972) and from surveys V-VII the special field guide for species identification prepared for Mozambique by Fischer et al. (1990). In surveys III through VII (1982–1990) records of the trawl stations and their catches are available in the NAN-SIS data bank.

Mainly because of bottom conditions most of the survey effort was expended in the Sofala Bank sector, with occasional extensions to other areas. Therefore, the majority of the results discussed below refer to the Sofala Bank only.

Acoustic estimates of pelagic fish

The degree of coverage for the acoustic investigations was estimated at 10 in survey V and 9 in survey VI. These surveys repeated the patterns of coverage from the previous surveys and the effort thus seems to have been adequate. Figure 5.4 shows, as an example, the survey routes and stations in the two surveys of the Sofala Bank in 1982.

Maps of the geographical distribution of small pelagic fish by integrator reading levels attributed to this group are available from surveys I, IV, V and VI. They show no recordings outside about 200 m depth and the highest densities were found at shallow depths (50 m and less), and mainly in the central part of the Sofala Bank between Quelimane and Beira.

Reliable estimates of the abundance of pelagic fish based on the echo integration technique are available from surveys II, III, IV, V and VI.

Figure 5.4 Survey routes and stations in the two 1982 surveys

As described in Section 3, the conversion factor C used in the acoustic method for estimating the biomass of fish from the observed integrator readings, was for survey II based on a target strength (TS) of -36 dB/kg (for 17 cm fish), for surveys III and IV on TS = -35 dB/kg and for surveys V and VI on TS = -34 dB/kg. In order to make the estimates comparable and since a TS level of -34 dB/kg now seems more appropriate, the biomass estimates reported from survey II have been reduced by 37% and those from surveys III and IV by 21%.

Table 5.14 Mozambique: Acoustic estimates of the biomass (t) of small pelagic fish adjusted to the same level of target strength, TS = -34 dB/kg

Table 5.14 shows the adjusted estimates. A considerable part of the variation is likely to have been seasonal or inter-annual changes in stock sizes. The simple mean is 158,000 t and using the proportions by families derived from surveys III and V (see below) the deeper shelf group of Carangidae and Scombridae represents about 100,000 t and the shallow water assemblage of Clupeidae and Engraulidae about 58,000 t. The biomass of the latter group is likely to be underestimated because of the probable distribution of these species into the inshore waters not covered by the surveys. This group also showed large variations in biomass, viz. 70,000 t in survey III and 120,000 t in survey IV. In survey VI the acoustic coverage was extended to include the shelf south of the Sofala Bank, from Bazaruto to the Boa Paz Bank. The estimated biomass of small pelagic fish in that area was 24,000 t.

Species composition of pelagic fish

Information on the composition of the pelagic fish is available from the fishing trials. The composition by families in catches from aimed fishing with the pelagic trawl in surveys III and V, (Table 5.15) showed a predominance of Carangidae, Clupeidae and Engraulidae, but the catches especially in the 1990 survey were small and the results are unlikely to be representative of the true relative abundance of the groups.

Table 5.15 Mozambique: Mean catch rates of pelagic families in aimed fishing with mid-water trawl (kg/h)

Pelagic fish formed also a substantial part of the catches in the bottom trawl as shown by the summary data of main groups in Table 5.16. The use of these more comprehensive data to analyse the composition of the group revealed (Table 5.17) a depth relationship as also discussed by Sætre and de Paula e Silva (1979) and Brinca et al. (1981): both surveys showed that Carangidae and Scombridae dominated the pelagic group in hauls deeper than 25 m, while Clupeidae and Engraulidae had the highest catch rates in the 10–25 m range.

Table 5.16 Sofala Bank: Mean catch rates of main species groups with the bottom trawl, 1982 and 1990 (kg/h)

Table 5.17 Sofala Bank: Catch rates of families of pelagic fish in bottom trawl by depth strata (kg/h)

The bottom trawl catch composition in the range 10–25 m depth is similar to that of the pelagic hauls, of which a major part was from the inner shallow part of the shelf. This indicates that the catch rates in the bottom trawl hauls reflected the abundance of the various families of pelagic fish over the shelf area covered by this range. Assuming that this was the case also for the deeper parts of the shelf and using the areas of the Sofala Bank shelf reported in Brinca et al. (1981), the estimates of the composition of small pelagic fish shown in Table 5.18 were obtained. The two sets of data agree in showing that Carangidae and Scombridae which inhabit the deeper parts of the Sofala Bank represented about two-thirds of the biomass of the small pelagic fish, while Clupeidae and Engraulidae found predominantly on the inner parts, amounted to about 25%. The latter families may be under-represented in these estimates because they can be expected to be abundant also on the shallow shelf inside the 10 m depth limit of the survey. The short-lived anchovies are likely to have an annual production cycle and this may not have been covered by the two surveys analysed here. In the coverages of September 1977, April-June 1978 and May-June 1982 about half the total biomass of small pelagic fish was reported to be anchovies (Stolephorus spp.).

Table 5.18 Sofala Bank: Composition of small pelagic fish estimated from catch rates in bottom trawl by depth ranges and shelf areas (%)

There is a large number of species of pelagic fish in the Mozambique fauna, but relatively few make up the main part of the biomass. The frequency of occurrence and the mean catch rates in bottom trawl for the dominating species of each family are shown in Table 5.19. Various Stolephorus spp. which occurred in about half of the pelagic hauls in each of the surveys (but represented only 16% and 3% of their total catches), should be added to the Engraulidae. With this addition the list of species shown in Table 5.19 is similar to the composition of main pelagic fish on the Sofala Bank reported from surveys I, II, and IV.

Table 5.19 Sofala Bank: Bottom trawl dominating species of pelagic fish, incidence (%) and mean catch rate (kg/h) by surveys

Biomass estimates of demersal fish

The bottom trawl data from surveys III (1982) and V (1990) give the best coverage of the Sofala Bank and are available in the NAN-SIS databank. These data were analysed further in this review.

The trawl data may be used for estimates of the standing biomass. In surveys III and V the trawl stations were distributed randomly by pre-defined strata. For the purpose of estimation of fish biomass from swept-area calculations simple depth strata have been used based on the depth configuration shown by Brinca et al. (1983) with the fish densities from Tables 5.21 and 5.22. The effective trawl width is as discussed in Section 2.1 assumed to have been 18.5 m and the catchability coefficient 1. The results are presented in Table 5.20. The estimates for the pelagic fish are as anticipated, considerably lower than those obtained by the echo integration method. The relationship between the two types of biomass estimates will vary between areas depending on fish behaviour and bottom depth, but the ratio found here is not unreasonable.

Table 5.20 Mozambique: Biomass estimates from swept-area trawl surveys (t)

The estimate of 52,000 t for the demersal fish in the two surveys is similar to that which can be calculated from the densities of demersal fish reported from survey II, 1980. Weighted mean densities by depth ranges from Table 2 in Brinca et al. (1981) gave a total biomass estimate of 54,000 t. From a series of surveys with various vessels in 1977–78 Sætre and de Paula e Silva (1979) reported a demersal fish biomass on the Sofala Bank based on the swept-area method of 67,000 t in summer and 36,000 t in winter (with a mean of 52,000 t).

Previous estimates of a demersal fish biomass on the Sofala Bank of 100,000–120,000 t (Sætre and de Paula e Silva, 1979, Brinca et al., 1983) were partly based on acoustics, partly on an assumption of an effective width of the trawl of only half of that used here.

Based on the simple assumption that the fish densities for the rest of the shelf of Mozambique inside the 100 m depth are the same as those estimated for the Sofala Bank, estimates for the whole shelf were obtained by raising those of the Sofala Bank by 44% resulting in about 220,000 t of pelagic fish and 97,000 t of demersal fish and other species.

Tables 5.20 and 5.21 show the densities by main families of demersal fish and other taxa and for depth strata, estimated from the mean catch rates. (Conversion to catch rates is obtained by multiplication by 30). Pelagic fish (defined as the families listed above) represented roughly half the catches, a level also reported from surveys I, II, IV and VI. Their proportion varies with depth and according to Sætre and de Paula e Silva (1979) also by season.

More than 40 families of demersal fish were identified in previous surveys (Brinca et al., 1981), but only five to ten families make up the main bulk of the catch. From those listed in Tables 9 and 10 in Brinca et al., 1981 and similar data reported from the other surveys it appears that goatfishes and lizardfishes are common over the whole shelf, while croakers, ponyfishes and grunts are typical for the shallow waters and threadfin breams and snappers for the deeper parts of the shelf.

The estimated mean density of pelagic fish was about the same in the two surveys, while the total for the demersal group appears to be about 30% higher in survey V. This is, however, only a sampling effect caused by a difference in the density of hauls in the various depth ranges. The estimates of biomass were, as will be shown later, the same for the two surveys.

The density of Cephalopoda was about five times higher in survey V than in survey III. This may be a seasonal effect, as the squids are known to have seasonal cycles in abundance. The other taxa appeared with about the same densities in the two surveys.

Table 5.21 Sofala Bank, survey III, 1982: Density of main demersal families and other taxa by depth ranges (t/nmi²)

Table 5.22 Sofala Bank, survey V, 1990: Density of main demersal families and other taxa by depth ranges (t/nmi²)

An important trawl fishery for penaeid shrimps takes place on the shallow parts of the Sofala Bank. The effort in this fishery was at a reduced level in a period in the 1970s, but increased from 1980 (FAO, 1981). The fish by-catch in the fishery was estimated at well over 20,000 t per year in the late 1980s. A substantial part of this was demersal fish of which croakers, grunts, goatfishes and ponyfishes formed important components (Brinca et al., 1983).

Changes in the shrimp fishery over the period covered by the surveys may have affected the composition of the demersal fauna. In order to determine if this is the case, pertinent data have been compiled in Table 5.22. This table shows the estimated total densities of the main demersal families in shallow parts of the Sofala Bank and the densities of the typical by-catch families for the surveys for which these data are available, viz., surveys II through VI.

Sætre and de Paula e Silva (1979) found a seasonal variation in the density of demersal fish in the shallow waters at the Sofala Bank with the mean density in summer, October to March double that of the winter, April to September. The high densities of survey II may thus have been a seasonal effect. An alternative explanation is that a reduced shrimp fishery in the 1970s had allowed the demersal stocks, especially croakers and grunts to expand, but with increased fishing in the early 1980s they were reduced to the relative stable levels shown by the subsequent surveys.

At the species level there is, as might be expected, more variability in the density between surveys III and V for which data are available, but there was, as shown in Table 5.23 a general consistency in the domination of each family by relatively few species.

Table 5.23 Sofala Bank, demersal fish, shallow water: Estimated densities of all main demersals and of “by-catch families” (t/nmi²)

In an analysis of demersal assemblages of fish on the East African continental shelf, Bianchi (1992a) found a fish fauna in Mozambique largely similar to that off Tanzania and Kenya. A homogeneous and well-defined group could be identified in shallow parts of the Sofala Bank, but in deeper waters the species represented groups similar to those found at corresponding depths in Tanzania and Kenya.

Table 5.24 Sofala Bank: Species composition of main demersal families in surveys III (1982) and V (1990). Proportion of total catch by weight of family (%)

The Delagoa Bay including the Boa Paz Bank was only covered with demersal trawl programmes in surveys I and V. Survey V showed an assemblage different from that of the Sofala Bank with dominance of sea breams and snappers in the depth range 25–50 m (Boa Paz Bank). Echosounder traces of semi-demersal fish were mostly found over rough stretches of bottom and it seemed therefore doubtful whether trawl catches provided unbiased observations of the composition and density of the demersal fish fauna (IMR, 1990b). (A similar bias will also have affected the observations further north since there are many parts of the Sofala Bank where rough bottom did prevent sampling by trawl.)

Review of findings

The Mozambique shelf was well researched in the period 1977–90 and the consistency of the main results confirms the adequacy of both the methods used and the execution of the programmes.

The standing stock of demersal fish was estimated at about 100,000 t for the entire shelf. There appeared to have been very little change in the mean density of this group over the period although some families at shallow depths, e.g., croakers and grunts, may have declined in the early 1980s as a result of increased effort in the shrimp fishery.

The variation of the biomass estimates of pelagic fish on the Sofala Bank between 100,000 and 210,000 t may well have been an effect of seasonal fluctuations in the stock of short-lived small pelagics.

It should be noted that neither the acoustic nor the trawl surveys covered the stocks of large pelagic fish, tunas and tuna-like species, which may be an important resource along the deeper shelf and in the adjacent waters outside.

The 1990 total catch was estimated at about 100,000 t, (source: DAP/SEP; IIP) well over half of which was shrimp by-catch and artisanal catch presumably consisting of small pelagic and small demersal fish. This might indicate a fair degree of utilization of the demersal potential, although surveys V and VI showed the presence of large-sized specimens of fish on hard bottom, seabreams, snappers and groupers suggesting a low rate of exploitation on these grounds.

The landings from directed fisheries for small pelagics: scads, mackerel, sardine and others were estimated at 10,000 to 15,000 t in 1990. With a standing stock estimate of small pelagics of more than 200,000 t there would seem to exist a considerable potential for increased utilization of this type of fish even if the group also form a considerable part of the shrimp by-catch and the artisanal catch.

Various attempts have been made by Mozambique's fishery administration to test the possibility of a higher utilization of this resource. A programme to determine if the stocks of small pelagics could be fished on an industrial scale was conducted in a joint activity between the Government of Mozambique and NORAD in 1985–87. A Norwegian fishing vessel of 21 m LOA (148 GRT) was used in a series of trial fishing experiments with purse-seine and mid-water trawl between August 1985 and September 1987. As reported by Sørensen et al. (1988), only low catch rates could be obtained with either gear although the fishing methods as such were judged to be effective. The failure of the trials were reported to be lack of areas with high densities of fish, although small schools and scattered fish were found. A commercial industrial fishery on these resources could thus not be recommended, but further trials were suggested with semi-pelagic pair trawls or high-opening bottom trawls on the Sofala Bank and small purse-seines and local boats in the Delagoa Bay.

This observation of lack of areas with high concentrations of pelagic fish is confirmed by an analysis of the distribution of acoustic observations of densities from surveys V and VI in 1990.

Table 5.25 shows the distribution of the biomass by ranges of density at which the pelagic fish was recorded. The general experience from other DR. FRIDTJOF NANSEN surveys is that only densities higher than 300 t/nmi² will represent areas where small pelagic fish aggregate in “school areas” suitable for commercial purse-seining. The data in Table 5.25 confirm the general impression obtained from the surveys of an absence of high density school areas suitable for purse-seining.

A fishery for scads and mackerel with bottom trawls was developed on the Sofala Bank and Boa Paz in the late 1970s and peaked in 1978 with a catch of 17,000 t and declined to 6,000– 7,000 t in the first half of the 1980s. In an analysis of the fishery by 1984–85 Gislason and Sousa (1989) recommended that fishing effort could be gradually increased. The acoustic estimates of biomass confirm a higher potential for these stocks.

Table 5.25 Sofala Bank: Distribution of biomass of small pelagic fish by density levels (%)

In general the productivity of the Mozambique shelf as measured from the estimated mean biomass of demersal and pelagic fish from the surveys over the period 1977–90 is at a moderate level for tropical areas with a mean density of 17 t/nmi². Limited areas of higher productivity may exist inshore. No information on the potentials of the open sea outside the shelf is available from these surveys.