10 SUMMARY REVIEW OF FINDINGS

10.1 REVIEW OF FAUNISTIC STUDIES

The data collected through the bottom trawl sampling programme on tropical continental shelves and slopes constitutes one of the most comprehensive databases on tropical demersal resources. This database is unique in that it includes data collected with the same type of gear throughout the programme and in being readily available for further analyses.

Data sets from the Western Indian Ocean, the Eastern and Western Central Atlantic and Eastern Central Pacific were selected to carry out a community study, i.e., detection of recurring patterns of species associations in the demersal trawl catches, correlation of these with the main environmental parameters and finally inference of different “communities” or “assemblages”. This study was a welcome contribution to the knowledge on species interactions and responses of different assemblages to environmental gradients, that is still limited despite the substantial literature on biology and stock assessment of tropical species. The results from this work may have applications in inferring unknown assemblages and catch composition in unexplored areas, and may be an aid to the stratification of fisheries by ecological regimes or in the identification of ecological regimes as a basis for holistic and multispecies modelling. For a detailed description of the results of the above work, the reader should refer to Bianchi (1991; 1992a, b, c, d; 1996).

The classification of communities, based on their predictability in time and space, describes the existence of a spectrum of community types ranging from deterministic to stochastic, of which the former are found in relatively undisturbed habitats, and the latter in environments so variable as not to allow predictable communities to develop. The type of species found will reflect this situation, with specialized feeders in the deterministic communities and nonspecialized feeders in the stochastic communities.

“Tropical” communities have generally been ascribed to the deterministic type because of the stable conditions found in tropical areas. Bianchi (op. cit.) has pointed to the inadequacy of this generalization as large areas of tropical shelves are subject to important seasonal fluctuations of different types. For example, many estuarine areas are flooded during the rainy season; the shallow and sharp thermoclines in shallow waters of the tropical west coasts of Africa and America are subject to vertical displacements due to tidal and internal waves, while shelf areas in the northwestern Indian Ocean are seasonally exposed to low-oxygen conditions due to upwelling. In such areas, according to ecological theory, more stochastic types of communities can be expected. On the other hand, the shelf waters off Tanzania, with rather stable conditions throughout the year, should show the existence of the more deterministic type of communities. This pattern clearly emerged from the community analyses, that also showed how highly adaptive species dominated the assemblages in the variable habitats.

The sharpest changes in species composition seem to occur along the depth gradient. This is particularly evident in shelf regions where different water layers impinge on the shelf bottom. Tropical shelves of the Eastern Atlantic and Eastern Pacific, characterized by the presence of a sharp and shallow thermocline separating the upper mixed layer from the much cooler deeper waters, belong to this category. Within the various depth strata, other factors - such as bottom type or influence by processes along the coast - become more relevant. However, zonation at around 30 m depth seemed to be present in all the areas studied, independent of the structure of the water column. The reason for the presence of this faunal boundary may indicate the separation between the shallow-water environment and the intermediate shelf environment and related differences in energy sources and flows. In shallow waters the relationship to the bottom must be stronger and primary production is enhanced by the nutrients brought by the rivers and those made available through vertical mixing.

One interesting finding from the above study was the dominance in shallow coastal waters of most regions of species of pelagic/semi-pelagic types. For example, shallow water communities of the Americas hold a rich mainly zooplanktivorous clupeoid fauna, mostly small (< 10 cm) and characterized by a rapid turnover. Now the question is whether this is the result of intensive fishing for shrimp, by bottom trawl with fine meshes or an ecological adaptation to an environment that can be rather unstable because of tidal movements, changes in turbidity, etc. The continuous variation would prevent this system from reaching a climax and favour phylogenetically primitive taxa (clupeoids). On the other hand, intensive fishing with bottom trawl and fine trawl meshes must lead to a selection of small clupeoids as compared to larger, longer-lived demersal fishes. The present situation probably reflects both mechanisms.

Communities of the deeper part of shelf also showed interesting adaptations. The relatively stable environments of the Western and Eastern Atlantic host communities of long-lived demersal species, i.e., lutjanids and sparids, rich in numbers of species and in abundance. In the Eastern Central Pacific, characterized by the presence of oxygen-depleted waters in the deeper part of shelf, the galatheid crustacean Pleuroncodes planiceps dominated, apparently undisturbed by any predation or competition. Off Pakistan, where the presence of water with low-oxygen concentrations is seasonal due to upwelling, some species adapted to this by being able to live throughout the entire water column (e.g., Trichiurus lepturus).

The above study showed that the marine assemblages found in the tropical region cannot be classified under a common denomination as found in the more general literature (e.g., Ursin, 1984). This is important because generalizations that may be valid for a particular tropical region do not necessarily apply to other regions at similar latitudes. Tropical seas (at least the areas covered) display a wide variety of combinations of oceanographic conditions, type of bottom and zoogeographic patterns, and the type of fauna reflects these conditions with a wide variety of forms and life strategies.

10.2 FISH ABUNDANCE AND ECOLOGICAL REGIMES

The data on fish abundance from the many different regions surveyed represent an important output from the programme as a whole. Since the same methods and fishing gear were used throughout, the biomass estimates are in principle comparable. The database provides a unique opportunity to compare predictions made in 1970 in “The fish resources of the ocean” (Gulland, 1970) with direct observations. The ecological regime of a region affects the productivity at lower trophic levels and hence to a large extent the fish abundance. One of the main mechanisms of productivity is the rate of renewal of nutrient depleted water in the surface, and the hydrographical descriptions of the survey regions focussed on this process. The classification of the regimes used below and in Table 10.1 roughly reflects rates of replacement of surface water.

An analysis of fish abundance by class of ecological system from all the DR. FRIDTJOF NANSEN surveys is attempted in Table 10.1. The density of small pelagic fish is the most important indicator since they are low in the food chain and their abundance may be expected to be most directly related to the primary production. Observations on demersal fish were included where available to complete the picture and also because in some surveys it was difficult to provide meaningful separate estimates of pelagic and demersal fish.

Table 10.1 Densities of biomass by types of ecological regimes for small pelagic and demersal fish in all areas and mesopelagic fish in the northwest Arabian Sea. Mean annual catches in survey periods included

The rate of exploitation which affects the standing biomass varied considerably between regions, and, in order to reduce the resulting bias, the mean annual catches were included in the density estimates.

By far the highest densities were found in the regions of perennial upwelling in the Canary and Benguela Current Systems off northwest and southwest Africa respectively. Information on the state of the stocks may explain some of the differences found between the two regions with considerably lower levels in the southern system. While the European sardine in the northern system has remained at a level of high production, the corresponding stock in the south, the Namibian pilchard, collapsed in the mid 1970s and has existed as a depleted stock since then. It may have been partly replaced by the horse mackerels, but those are higher in the food chain and less efficient utilizers of biological production. The existence of components of the fauna of small pelagics which were not included in the biomass estimates may also complicate the comparison. Observations indicate that one such component, myctophids, were less abundant off Morocco-West Sahara than in Namibia where these fish often complicated the estimates of horse mackerels. Pelagic gobies also had a high abundance near the southern upwelling centre in Namibia.

In the regions of seasonal upwelling in the eastern boundary currents, for example, Mauritania to Guinea Bissau and Angola the fish density was only about one-third of those in the regions with perennial upwelling.

For the open ocean upwelling in the western Arabian sea, the levels given correspond to total stocks estimates of 33–52 million t of mesopelagic fish.

Among the other seasonal upwellings those associated with the southwest monsoon cause the high production off Oman and Somalia. The relatively lower density off Somalia may be an effect of the very narrow shelf here. There is a clear seasonal effect of the southwest monsoon on the hydrographical regime also in southwest India along the Malabar coast. Sri Lanka is affected by both monsoons and consequently has a rather high density, but Myanmar only by the northeast monsoon and has rather low fish densities.

The surveys apparently did not detect any effect on fish abundance of the upwelling along the north coast of Mozambique during the northeast monsoon, which perhaps may be explained by the very narrow shelf along this coast.

The Caribbean coast of Venezuela offers an impressive example of the effect of upwelling on fish production. The density of small pelagics along the coast of the Oriente province where upwelling is intensive was estimated at 116 t/nmi², while it was only 10 t/nmi² for the neighbouring coast further west where the Caribbean Current has shifted offshore. Still further west, a local wind-induced upwelling was again observed to cause increased fish density off the Guajira Peninsula.

With the exception of Oman, Somalia and Venezuela-Oriente, the fish densities in these various regimes of seasonal upwelling were generally at an intermediate level, similar to those of the seasonal upwellings in the eastern boundary current systems.

The surveys of Pakistan and the Suriname-Guyana coast covered systems where current dynamics seasonally lifts deeper water onto the shelf and the surface layers are enriched by vertical mixing. The resulting increase of fish density was especially evident in Pakistan.

In the tropical regions of year-round stability of the surface layers the observed fish densities were low, on average about one-fifth of those of the seasonal upwellings. Obviously the surveys did not include coral reefs and mangrove areas, which are known to be highly productive.

It would have been desirable to use a more quantitative classification of the hydrographical regimes with, for instance, measures of duration and intensity of the process of renewal of surface water. Data on fish composition, the state of exploitation of the stocks, fish age, etc., would also be required for a more comprehensive study of fish production by type of ecosystem. However, the relatively simple analysis made here clearly demonstrates the limits to fish production set by the types of hydrographical regimes.

The use of biomass estimates per unit coastline instead of unit shelf gives different relationships of indices of fish production by ecosystems. In northwest Africa (Table 6.??) fish density in the area of perennial upwelling from Cape Bojador to Cape Blanc is 1.8 times higher than in the northern area from Cape Safi to Cape Bojador when measured by unit shelf area, but 3.3 times higher if measured by unit coastline length. In the region from Mauritania to Guinea Bissau fish density by unit shelf is only half that of Cape Safi to Cape Bojador, but nearly the same if measured by unit coastline. This difference may partly be caused by the effect of shelf width on fish production which is also indicated by some of the observations in the Indian Ocean, comparison of the northeast Somalia system with that of Oman, and the upwelling off northern Mozambique. The fish density by unit shelf area is of greatest general interest, but the difference between these indices may help to explain important features of the ecosystems.

10.3 REVIEW OF BIOMASS ESTIMATES BY REGIONS AND COUNTRIES

The biomass estimates obtained from the DR. FRIDTJOF NANSEN surveys will be reviewed below, following the sequence of the previous chapters, and where possible the survey results will be compared with one of the first estimates of the world's marine fish resource potentials. For this purpose rounded summary data have been extracted from Gulland (1970). Where possible these data have been converted into units of t/nmi². It should be noted, however, that Gulland's figures represent potential catches, while the results of the surveys represent the standing stock. The so-called Gulland formula, MSY = XMB_o, where X is usually 0.5, M is the natural mortality coefficient and B_o is the unfished biomass or Cadima's formula, MSY = 0.5 (Y + M * B), where Y is the total catch in a year and B is the average biomass in the same year (Sparre and Venema, 1991). These and similar formulas can be used to convert the results of surveys into estimates of potentials and vice versa.

10.3.1 The Arabian Sea and adjacent Gulfs

The North Arabian Sea with its seasonal systems of high primary productivity represented a region for which the Indian Ocean Programme had held the highest expectations to potentials for development. The information provided by the main resource surveys mounted by the programme was presented and reviewed at a special workshop in Karachi early in 1978 (FAO/IOP, 1978). There is historical interest in recording the gains in knowledge which may be attributed to these first survey efforts and in relating the new information to the main features of the ecosystem of this ocean.

The areas of high productivity were related to the upwellings associated with the southwest monsoon. Cushing (1971a and b) describes three coastal upwellings, in the Somali Current, southwest off the Arabian peninsula and off the Malabar coast. His maps of estimated tertiary production show high levels extending more than 500 nmi seaward and covering nearly half the Arabian Sea. This is referred to as high productivity in associated oceanic divergences or possibly the biological effects of the coastal upwelling drift seawards in the swift current. The boundary between the true coastal upwelling and the offshore divergences was described as being less precisely defined in the northwest Arabian Sea than in the four major eastern boundary current upwellings.

In Gulland (1970), the Arabian Sea was recognized as a highly productive area with average potentials of 13.4 t/nmi² of shelf area (0–200 m) for demersal fish and 16.8 t/nmi² of pelagic fish (Table 10.2). However, compared to the estimates for areas like the southeast Atlantic and the Eastern Central Pacific, the estimated densities were very modest, although the total potential was high. The large resources of mesopelagic fish, which absorb a large part of the estimated potential had not yet been identified. It is obvious that the DR. FRIDTJOF NANSEN surveys have very much contributed to a better understanding of fish distribution and population dynamics in this area.

Table 10.2 Arabian Sea: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

A review of the DR. FRIDTJOF NANSEN estimates of standing biomass and densities of the pelagic and semi-demersal fish in the various systems is given in Table 10.3 which also includes the mesopelagics on the slope and adjacent oceanic parts of the northwest Arabian Sea. The densities of pelagics and semi-demersals are highest in the western coastal upwellings, intermediate on the Malabar coast of India and lowest on the Pakistan shelf. The best estimate from the 1975–77 surveys of total biomass of small pelagic and benthopelagic fish in the main shelf regions of the Arabian Sea is thus 6.3 million t. (The levels reported at the 1978 Karachi Workshop (FAO/IOP, 1978) for the area from Somalia to Pakistan were, however, only about half of those shown in Table 10.3, because the target strength used in the first surveys is now judged as having been too high.)

Table 10.3 Arabian Sea: Summary of results of acoustic surveys in the 1970s. Estimated mean standing biomass and densities of small pelagic, semi-demersal and mesopelagic fish by areas

The reported catch of small pelagic and semi-demersal fish from the Western Indian Ocean for 1976 was about 1.2 million t (FAO Yearb. Fish. Stat., 45). Most of this would be from the surveyed area, and consequently the survey results, which for the Karachi workshop were interpreted to indicate a standing biomass of these fish of only about 3 million t, did not show any notable potential for expansion of the fisheries based on conventional shelf resources. The high abundance of mesopelagic fish in offshore and oceanic parts of the ocean, reported as about 100 million t at the Karachi Workshop caused considerable attention and interest, but it was recognized that various technological and economic problems might prevent an immediate utilization of this resource and that further research and development work would be required.

The main effect of these first surveys was thus in general to correct and perhaps over-correct the vision of the Arabian Sea as a region with a new great fishing potential. This vision was as discussed in Section 3.1, based on a widely accepted concept of the Arabian Sea as “one of the more productive parts of the world oceans” in terms of primary productivity and on a comparison of catch per area of the Indian Ocean with such data for the Atlantic and the Pacific.

It should be noted that in one of the background studies for the IOP, Cushing (1971b) presented estimates of tertiary production in the western coastal upwellings, Somalia and South Arabia of only 3.3 million t, a level which fits the biomass estimates for the Arabian Sea shown in Table 10.3.

All descriptions based on the findings of the IIOE relating to production in the Arabian Sea emphasised, however, its great seaward extent. In the early more general considerations of the high productivity at low trophic levels of the Arabian Sea and when assessing its significance for the fishery potentials of this ocean insufficient distinction seems to have been made between the coastal and the oceanic ecosystems. In a later description of upwelling in the Arabian Sea (Luther, 1991) such distinction is clearly made. The process is said to be driven primarily through the mechanism of open-ocean upwelling forced by the wind stress curl associated with the atmospheric Findlater Jet over the Arabian Sea during the southwest monsoon. This open ocean upwelling is combined with a narrower band of coastal upwelling due to the alongshore components of the winds.

Variability in the open ocean upwelling on an interannual and decadal scale can be related to that of monsoon rainfall in India. The data indicate a strong monsoon year in 1975 and an average one in 1976, the two years of the DR. FRIDTJOF NANSEN surveys. In view of the short lifespan of both the mesopelagic and the neritic small pelagic fish, variability in the extent and intensity of upwelling is likely to cause variability in standing stock biomass.

In a recent review of available information on zooplankton and nekton in the Arabian Sea, Peterson (1991) summarizes data on zooplankton abundance which shows biomass densities in productive parts of the open Arabian Sea of about 4 g dry weight m^-2, the same order as in the Somalia coastal area and similar to levels observed in eastern boundary current upwellings. In terms of fishery potentials this picture of the Arabian Sea as a highly productive region must be qualified by a faunistic distinction between the shelf and the oceanic provinces. The early acoustic surveys, the UNDP/FAO Pelagic Fishery Project on the Malabar Coast of India and the 1975–77 DR. FRIDTJOF NANSEN surveys provided important information for such a distinction. As shown in Table 10.3 the small pelagic and semi-demersal fish of the continental shelves around the Arabian Sea was estimated to have a standing biomass, corresponding roughly to the annual production, of 6.3 million t. This is probably an underestimate of the total production since the species and area coverage is not complete.

Figure 10.1 Satellite images showing phytoplankton concentrations in pre-and post-southwest monsoon conditions, May-June left and September-October right. Image prepared by the Goddard Space Flight Centre and the University of Miami. Source: Olson, 1991

The mesopelagic fish of the oceanic province of the western part of the ocean was estimated to have a standing biomass of 33–52 million t. In view of the short lifespan of these fish the annual production may exceed the mean standing biomass. The surveys provided no information on the oceanic fish of higher trophic levels, tunas, etc. The highest densities of mesopelagics were still found near the shelf edge and slope, e.g., 190 t/nmi² in the Gulf of Oman compared with 75–120 t/nmi² in the offshore areas. But even if the production of mesopelagics near the slope is considered as “shelf production”, the aggregate represents only about 20% of the total.

These surveys thus confirmed that the Arabian Sea is a region of high fish production, but a high proportion (about 80%) is mesopelagic fish produced off the shelf, in the slope and adjacent oceanic waters. That this distribution of production at higher trophic levels to a large extent reflects the geographical distribution of primary production is demonstrated by Figure 10.1 which shows a satellite image of the phytoplankton concentrations in two seasons, May-June and September-October.

The shelf hydrography in the northwest Arabian Sea is strongly affected by the southwest monsoon when oxygen-deficient bottom-water intrudes onto the shelf to shallow depths and affects fish distribution. Some enhancement of the productivity also occurs on the Pakistan shelf through shorewards advection and lifting of nutrient-rich deeper water by the east-and south flowing coastal current set up during the southwest monsoon. Furthermore, there is the well known strong coastal upwelling in the west, off Somalia, Yemen and Oman.

Pakistan, Iran, Oman, Yemen and Somalia had been covered in the initial exploratory programme in 1975–76. From 1977 to 1984 these surveys were repeated to confirm previous findings and obtain more detailed and comprehensive descriptions of the resources.

Pakistan

The estimates of pelagic fish were 550,000, 450,000 and 800,000 t in the three surveys compared with a mean of 750,000 t in the five 1975–76 surveys. Table 10.4 shows the estimates of the mean standing biomass from the 1983–84 surveys.

Table 10.4 Arabian Sea: Summary of survey results in 1983–84. Estimated mean standing biomass and densities of small pelagic and demersal fish by country

It seems unlikely that the semi-demersal fish for which there is an acoustic biomass estimate of 140,000 t was fully represented in the bottom trawl catches, and the “best estimate” of the mean standing biomass of demersal fish is perhaps 300–350,000 t.

The observed biomass represented partly exploited stocks. The reported landings of marine fish increased from 261,000 t in 1981 to 622,000 t in 1993.

Oman

The biomass estimates of the small pelagic fish were 1.0, 1.3 and 1.4 million t in the 1983–84 surveys compared with 250,000 t in early 1975 and 1.5–2.6 million t in the four subsequent surveys from late 1975 to late 1976. There has probably been a negative bias in the first two of the three 1983/84 surveys caused by instrument saturation.

The best estimate of demersal fish was 390,000 t when a November 1983 figure of 260,000 t was rejected as being caused by a seasonal low availability in deeper waters. This may still represent an underestimate as acoustic assessment of semi-demersal fish ranged from 38,000 to 148,000 t in the three surveys.

Table 10.4 shows the mean estimates of standing biomass and densities from the 1983–84 surveys. The resource of mesopelagic fish in the Gulf of Oman with a biomass estimated at 5 million t is described in Section 3.2.2.

A set of comparable data on Oman's fish resources are available from the 1989–90 surveys with the RASTRELLIGER which showed 470,000 t of pelagic fish and 500,000 t of demersals.

The main difference between the findings of the two sets of surveys is the low abundance of pelagic fish in 1989–90 compared with 1983–84. As noted above, a similar observation of low abundance of pelagic was made in the first 1975 survey, while pelagic fish were found in high abundance in the four subsequent surveys from late 1975 to late 1976. Such fluctuations indicate a dependence of populations size on a variable environment and interannual variations in the intensity and duration of the southwest monsoon are well known and reflected in the Indian summer monsoon indices (ISMR) available since 1871. The ISMR index for 1987 showed this to be the fourth driest year since 1871. The weak monsoon of 1987 could thus have caused the low biomass of pelagic fish in 1989–90 found in the RASTRELLIGER surveys.

The total landings of Oman fisheries have varied between 100,000 and 160,000 t since 1980. The high abundance shown in the surveys of such species as Japanese threadfin bream and Arabian and Indian scads is not reflected in the landings, probably due to the small size of these fish. There exists a large potential for growth of the fisheries in Oman.

Yemen and Northeast Somalia

The two surveys, from the northwest and southwest monsoon seasons respectively, showed the well known effect in this part of the Indian Ocean of redistribution of demersal fish caused by upwelling of oxygen depleted water during the southwest monsoon.

Using in each case the surveys assumed to have given the best coverage of the target group, the estimates of standing biomass shown in Table 10.4 were obtained.

A purse-seine fishery for oil sardine on the Yemen coast ceased about 1980. There is a considerable history of attempts and efforts to develop marine fisheries in Somalia few of which met with any success despite the relatively promising baseline resource studies. The problems of utilizing the marine fish production in Somalia are no doubt related to such factors as the remoteness of the northeastern coast, the limited tradition in fisheries, the narrow shelf, the rough sea and wind conditions during the monsoon and the low fish consumption in the country.

10.3.2 Eastern Indian Ocean

Relatively high potentials were assigned to several areas by Gulland (1970), see Table 10.5. In particular the high potential of demersal fish assigned to the shelf of Indonesia draws attention.

Table 10.5 Eastern Indian Ocean: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

Large parts of the shelf in the northern part of the Eastern Indian Ocean were surveyed in 1978–80 after the completion of the first programme in the Arabian Sea. The first-generation echosounder was still in use and the acoustic estimates of dense-schooling fish may be too low. The swept-area trawl method had still not become a routine part of the programme, and except for the Bangladesh assignment the demersal and semi-demersal fish were mostly assessed by acoustics which again indicates underestimation. Semi-demersal fish was, however, the main component of the biomass in Sri Lanka and was also abundant in Myanmar.

In order to cover different monsoon seasons and verify findings Sri Lanka's shelf was surveyed three times and those of Bangladesh and Myanmar twice. Relatively large parts of the inshore shelf in these areas were too shallow for surveying: in Bangladesh more than 40%, in Sri Lanka with Palk Bay and Strait also about 40%, but in Myanmar only about 7% in the Delta area.

The hydrography of the shelf waters is also affected in these eastern parts by the monsoons, but in different ways in the various areas. Lifting of the transition layer and intrusion of oxygen-deficient water onto the shelf was found to occur on the west and southwest coast of Sri Lanka during the southwest monsoon, but on Sri Lanka's northeast coast and on the Bangladesh and Myanmar shelf during the northeast monsoon. These processes had distinct affects on the depth distribution of demersal fish - especially in Bangladesh and Myanmar - and enhance productivity in the surface layers, but widespread upwelling was not observed. River run-offs affected the surface salinity over wide inshore parts in Myanmar and Bangladesh in the post-southwest monsoon season.

Sri Lanka

The total biomass on the west, south and east shelves was found to be 400,000–500,000 t with some seasonal variation. This represents a mean density over the surveyed shelf of 67 t/nmi². The most important component of this was demersal and semi-demersal fish assessed at 250,000– 350,000 t and with a potential total yield of 50,000–70,000 t. Some 75% of these resources was located on the west and south coasts (see Table 10.6).

Sri Lanka's reported marine landings increased from about 150,000 t in 1979 to about 220,000 t in 1993. There is some uncertainty regarding the total potential since the wide northern shelf could not be surveyed.

Bangladesh

The total marine fishery resources of Bangladesh was estimated at 280,000 t, of which 130,000 t each of pelagic and demersal fish, 16,000 t of sharks, rays, cephalopods, etc., and 5,000 t of deep water fish. This represents a density of 17 t/nmi² over the shelf to 100 m depth. The reported marine fish landings of Bangladesh were 100,000 t in 1980 and 133,000 t in 1993.

Myanmar

The estimated biomasses of small pelagic and demersal fish are shown in Table 10.6. There is a strong seasonal variation that could partly be an effect of seasonal production of engraulids and aggregation from the deeper shelf caused by the intruding low-oxygen water.

Densities of fish measured as biomass per unit shelf area showed means of 12 t/nmi² in September-November and 24 t/nmi² in March-April. These represented low to moderate levels, indicating only a moderate effect on the total production from the upwelling caused by the northeast monsoon and by the considerable river discharges. For comparison, a similar estimate of mean density of biomass measured by acoustics on the Malabar shelf (see Section 3.2.1) with its upwelling from May to September during the southwest monsoon, was 67 t/nmi².

Table 10.6 Eastern Indian Ocean: Summary of estimates of biomass and densities by country

Taking the simple mean of the two biomass estimates as mean standing stock and yield fractions of 0.5 and 0.25 for small pelagic and semi-demersal fish respectively, a theoretical potential annual yield of about 470,000 t is obtained. Since the stocks were already exploited (the 1979 landings were reported to be 400,000 tonnes), the total potential would be higher probably some 600,000 t.

Thailand, Malaysia and Sumatra

Parts of the shelfs off western Thailand, peninsular Malaysia and northern Sumatra were covered only once implying a character of exploration, but each survey was conducted with a fair effort of trawling and acoustics.

In this tropical region a stable surface layer was found to cover a large part or even the whole shelf. Thus the thermocline was found at 40–50 m depth on the 80 m deep shelf of the east coast of peninsular Malaysia, at 40 m on the 100 m west coast shelf, and the surface layer covered the whole of the 100 m shelves of the west coast of Thailand and the north and west coast of Sumatra. As indicated by the catch rates by depth ranges, the main part of the fish was found within the stable surface layer: a mean of 87% of the pelagics and 83% of the demersals were found above the thermocline.

The very high species diversity known from the region was confirmed. Up to 160 species from 80 families were identified in one survey with 20 species of carangids only.

The biomass estimates are shown with shelf densities in Table 10.6. The densities are low and remarkably similar, especially for the demersal fish. These were single surveys made from June through August and a seasonal variation would not be observable. Still, it is thought that the densities found are roughly representative for these stable tropical systems.

10.3.3 Southwest Indian Ocean

The northern part of this area has generally been considered to have a low potential, as shown in Table 10.7, from Gulland (1970). This is largely confirmed by the DR. FRIDTJOF NANSEN surveys.

Table 10.7 Southwest Indian Ocean: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

In the early 1980s, there was a great interest in establishing the resource potentials of Kenya, Tanzania and Mozambique. The quite considerable survey programme included four surveys in Kenya between 1980 and 1983, three surveys in Tanzania in 1982–83 and five survey programmes in Mozambique over the period 1977–90. The latter were part of an extensive bilateral programme of fisheries development between Norway and Mozambique. Through these many assignments the area was very well covered with acoustic and swept-area trawl surveys at different seasons.

During the southwest monsoon the current along the East African coast is directed northward from latitude 10°S and with the resulting onshore movement of the surface water off Kenya and Tanzania there is no upwelling. During the northeast monsoon wind-induced coastal upwelling occurs off the narrow shelf of northern Mozambique down to Angoche at latitude 16°S. The shelves have in general tropical conditions with stable surface layers. River runoffs affect the surface salinity of inshore parts seasonally and may also cause some local enrichment of productivity.

On the narrow Kenyan shelf the main part of the biomass was small pelagic fish and ponyfish, while in Tanzania and Mozambique demersal fish represented a larger part of the biomass with about one-third of the total.

Table 10.8 shows the summaries of mean standing biomass in the survey periods and the densities over the shelf to 200 m. There are uncertainties concerning shallow inshore parts which could not be covered. The densities are largely similar and their low levels are more or less as expected from the low-productive ecosystems. The reported landings may be underestimated due to inadequate statistical systems for the small-scale fisheries.

Table 10.8 East Africa: Review of biomass estimates, densities and reported landings

10.3.4 Eastern Central Atlantic

This large area that includes one of the world's most important upwelling areas was assigned a relatively low potential in Gulland (1970) (Table 10.9). The DR. FRIDTJOF NANSEN surveys proved that this is one of the most productive areas as far as small pelagic fish are concerned.

Table 10.9 Eastern Central Atlantic: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

The survey programme on the shelf of Northwest Africa as a whole consisted of a discontinuous series of investigations. There were four assignments, but with considerably different coverage and duration: 11 months in 1981–82, 4 months in 1986, 1 month in 1989 and 3 months in 1992. The area from Morocco to Sierra Leone was surveyed three to four times, while the southern area from Côte d'Ivoire to Ghana was surveyed only twice. The main objective in all surveys was to describe the distribution, composition and abundance of the small pelagic fish. Demersal fish were investigated by trawl surveys in a few selected areas only.

Strong seasonal variability and large contrasts between the waters in the north and in the south are the main characteristics of the Canary Current upwelling system. The latitudinal shift in upwelling results in a coastal surface temperature front in the southern part of the system between cold upwelled water to the north and warm southern tropical water. This front is found to the north near Cape Blanc in August-September and south of the Sherbro Islands in March. Between approximately Cap Timeris and Cap Bojador upwelling is most intensive and prevails throughout the year. Southward to Guinea, at about 10°N, upwelling is seasonal and takes place during late winter and spring.

Cape Blanc devides the region into a northern temperate and a southern sub-tropical regime, creating two main sub-regions for the assemblages of small pelagics: to the north mainly temperate species, European sardine, Atlantic horse mackerel and chub mackerel, and southward mainly tropical, round and flat sardinella, Cunene horse mackerel, yellow scad and other carangids and triggerfish (Balistes capriscus).

The surveys were organized by sub-regions and the main pelagic stocks covered by these subregional surveys were roughly as follows:

Cape Safi-Cape Blanc: Sardine stocks and main part of stocks of Atlantic horse mackerel and chub mackerel

Mauritania-Guinea Bissau: Main part of round and flat sardinella stocks, Cunene horse mackerel, scads and other carangids and triggerfish in the south.

Guinea-Sierra Leone: Triggerfish, sardinellas, scads and other carangids.

Côte d'Ivoire-Ghana: Sardinellas, Cunene horse mackerel and triggerfish.

Cape Safi to Cape Blanc

The mean estimates of standing biomass of the main groups in this area are shown in Table 10.10 for the 1986, 1989 and 1992 surveys which are thought to have given more reliable and complete data than the 1981–82 surveys. The data which include the mean annual reported landings in the period, are shown by the shelf regions Cape Safi to Cape Bojador and Cape Bojador to Cape Blanc which are thought to hold different sub-stocks of sardine. The density estimates used are based on combined biomass and catches which are thought to give better indices of the fish productivity of the systems than using standing biomass only.

Table 10.10 Cape Safi (Morocco) to Guinea Bissau: Mean estimates of standing biomass and densities of small pelagic fish from surveys in 1986, 1989 and 1992 and mean annual reported landings in the same period

A comparison of biomass with the level of the landings indicates that the central sardine stock in the sub-region Cape Safi to Cape Bojador was intensively fished in the period. The standing stock in this sub-region varied between 320,000 t and 1,650,000 t. The southern stock must only have been lightly fished and the stock estimates varied only between 3 and 4 million t.

It is thought that the surveys did not provide good estimates of biomass of horse mackerel and chub mackerel. For the Atlantic horse mackerel incomplete coverage of aggregations in the shelf slope and a possible oceanic distribution may have caused an underestimate.

The region Cape Bojador to Cape Blanc where upwelling is perennial had by far the highest density with nearly 300 t/nmi², nearly double that of the northern region Cape Safi to Cape Bojador.

Mauritania to Guinea Bissau

There were three highly variable estimates of the sardinellas from the 1981–82 surveys, which were probably caused by inadequate coverage of the distribution in shallow inshore parts. The likely time series is 500,000 t in 1981–82, 700,000 t in 1986 and the surprisingly high level of 4 million t in 1992. The estimates of standing stocks of carangids which consisted mainly of Cunene horse mackerel and yellow scad were approximately 900,000 t in all surveys.

Table 10.10 shows the mean estimates of the biomass over the period of the surveys and the mean annual reported landings. The reported landings of both sardinella species were low in the early 1980s, about 150,000 t, but increased to well over 200,000 t by 1986 and exceeded 300,000 t in the early 1990s. The landings of horse mackerels, etc., were low compared with the standing stocks.

The densities were here as expected much lower than in the region of high upwelling further north, about one quarter of those found between Cape Bojador and Cape Blanc.

Triggerfish

The occurrence of the grey triggerfish (Balistes capriscus) in high abundance in the 1970s and 1980s in two regions, Ghana-Côte d'Ivoire and Sierra Leone-Guinea Bissau was probably an unusual phenomenon. Most triggerfishes are slow-moving solitary reef-dwellers and this seems to be the more normal behaviour and habitat also of the grey triggerfish. However, the stocks in the Eastern Central Atlantic were able to expand their populations to a remarkable and (as would appear) unusual size, probably by utilizing the relatively high productivity of the yet tropical regimes of the shelf waters of the Western Gulf of Guinea and the coastal seasonal upwelling system south of Senegal. The DR. FRIDTJOF NANSEN surveys 1981–82 coincided with the culmination of the triggerfish stocks and those of 1986 and 1989 with their decline and collapse.

Biomass estimates of triggerfish from the relevant DR. FRIDTJOF NANSEN surveys were shown in Table 6.19 together with available data from other similar surveys. For the western stock there seems to have been a rapid growth from about 0.4 million t in 1978–79 to about 1.4 million t in 1982 and with a decline to only 0.2 million t in 1986. The eastern stock was about 0.5 million t in 1981, half that of the western stock, and had declined to 140,000 t by 1986 and to virtually zero in 1989.

The stocks were fished by Ghana and the USSR between 1972 and 1990 with an accumulated total yield for the period of well over 0.5 million t. It seems unlikely that the collapse was caused by fishing pressure and the growth and decline of the stocks is probably a phenomenon of natural ecosystem variability.

10.3.5 Southeast Atlantic

The potential of the resources in this area was already fairly well known in the late 1960s. Gulland's (1970) estimates of the potential catches for the whole region were 1,305,000 t demersal and 3,900,000 t pelagic fish. This estimate incorporates South Africa and a part of the Indian Ocean. For Angola plus Namibia the figures are for demersal fish 945,000 t and for pelagic fish 2.6 million t (see Table 10.11).

Table 10.11 Southeast Atlantic: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

In Table 10.12 the average results of the surveys with the DR. FRIDTJOF NANSEN are presented. The total biomass or average standing stock in Angola plus Namibia is 647,000 t of offshore demersals and 3,6 million t of pelagics. The survey results alone would indicate a lower potential than that given by Gulland, but in this area the extremely heavy fishing pressure in the 1980s should also be taken into account. It may be concluded that Gulland's prediction was reasonably accurate in this case.

Table 10.12 Namibia-Angola: Mean estimates of standing biomass and densities of small pelagic an demersal fish from surveys, and mean annual reported landings in the survey period

The programme in Angola was very extensive: 12 surveys between January 1985 and September 1992 with both small pelagics and demersal fish as main objectives. Deep-water shrimps and other resources in the slope were also investigated. The stocks and their environment were well described in the 1985/86 surveys, and the subsequent assignments had partly a monitoring character.

The Angola shelf forms a northern extension of the Benguela Current System: seasonal upwelling in winter spring and sub-tropical conditions in summer with poleward surface current.

The division of the Benguela Current System into a southern temperate and a northern subtropical regime is found off southern Angola. Here sardine, Cape horse mackerel, large-eye dentex and other species represent shared stocks with Namibia, while northwards sardinellas and Cunene horse mackerel are the important pelagic fish partly shared with Congo and Gabon.

The many surveys produced a wealth of information on the resources by season and subregion. A summary of estimates of mean standing biomass of the main groups is shown in Table 10.12. Time-series of the sardinella stocks showed a trend of decrease from 1985 to 1989 possibly caused by high fishing rates, but with a stock recovery in the early 1990s when fishing declined.

There was a trend of decline of indices of deep-water shrimp abundance over the period of the surveys. A similar trend was observed in the Benguela hake which forms a by-catch in the shrimp fishery.

A sharp decline in the fishery on the large-eye dentex in the Cunene-Tombua region in the late 1980s may have caused the high biomass estimates for that species found in the last surveys.

The mean standing biomass of sardinellas and Cunene horse mackerel of 500,000 t compares with a mean annual catch of 230,000 t in the period which perhaps roughly indicates that these resources were fully utilized during that time. Data on fishing for Cape horse mackerel in Angola are not readily available, but a full utilization of that stock and of a recovered stock of sardine would raise Angola's potential annual yield of small pelagics to 300,000–400,000 t. This fits with the catch levels in the purse-seine fishery of the former fish meal industry in the 1950s and 1960s. Angola's reported catches of demersal fish in 1991–92 were, according to the official statistics, less than 20,000 t. It is uncertain whether these data include all fisheries, but in general there seem to be resources available for expansion of the national fisheries in many directions. The deep-water shrimps seem to have been too heavily fished.

The DR. FRIDTJOF NANSEN survey programme in Namibia 1990 through 1993 started immediately prior to the country's achieving Independence and consisted of bi-annual separate surveys of the demersal and pelagic stocks. The fisheries in Namibian waters prior to the declaration of an EEZ at Independence was international and had a considerable history with well-known stocks and yields. Substantial knowledge of the sea and the stocks was acquired through research organized in the period 1970–90 by ICSEAF, but Namibia took over a legacy of largely depleted stocks.

Upwelling and the intrusion of oxygen-deficient water onto the shelf are the two most important features of fishery oceanography. Upwelling is perennial off Namibia with two major centres, one in the north near Cape Frio and the other in the Lüderitz region in the south. Anoxic conditions were often observed in an inshore belt off the central part. Mass fish mortalities occur in this region.

The main pelagic stocks are pilchard with some round herring and anchovy inshore and Cape horse mackerel offshore. Cape and deep-water hakes are the main demersal stocks.

Estimates of mean standing biomass of the pelagic stocks in the survey period are shown in Table 10.12. The pilchard stock which in the 1960s sustained annual yields of more than 0.5 million t collapsed in the early and mid 1970s and catches in the 1980s were maintained at 50,000–70,000 t despite the clear evidence of the collapse of the stock caused by overfishing.

The stock level of pilchard was estimated at 600,000 to 700,000 t from 1990 to 1992, but declined to about half that level in 1993. The observed decline in 1993 coincided with a northward shift of its distribution, and the estimates include pilchard found in southern Namibia. A possible explanation may be found in the structure of the original Namibian pilchard stock. As a whole it may have represented a super-population consisting of several partly mixing sub-populations arranged along the coast. Stocks in the central area would be more vulnerable to fishing and to predation from the seal population. The natural preypredator relationship between the two populations has been distorted by the collapse of the pilchard stock and at its present low level the pilchard stock is not sustainable.

The anchovy stock was also overfished by the early 1980s and no lasting recovery has occurred.

The Cape horse mackerel stock probably had a mean standing biomass of about 2 million t in the survey period and the reported landings indicate that there may have been an unused potential catch.

The history of the hake fisheries showed total annual landings of 0.5–0.8 million t over a nearly 10-year period up to 1976 and then a decline with landings generally at 250,000– 350,000 t. The stocks were very low in 1990, and the first survey showed a standing biomass of the Cape hake stock of only 0.5 million t, of which only 100,000 t was fishable. Namibia introduced a strict management regime, and estimates of the standing biomass increased by about four times over the first two years as shown in Figure 7.?. The expansion then stopped and there was some decline in 1993.

This break in the recovery of the stock was probably due to insufficient recruitment. Recruitment variations in the Cape hake prove to be high, similar to many gadoid stocks in the North Atlantic. Whereas in 1993 recruitment has been assessed at a level of 2 billion fish, in previous periods, 1968–74 and 1982–85, estimates indicate that the mean recruitment was about 4 billion and the abundant cohorts comprised more than 6 billion fish. The 1991 yearclass was assessed at about 4 billion in November 1992, but at only 2.5 billion in February 1993, demonstrating a phenomenon of mass mortality in the region off Walvis Bay.

10.3.6 Eastern Central Pacific

The DR. FRIDTJOF NANSEN surveys covered a large part of the tropical section of this region, for which Gulland (1970) had estimated potentials of 1 million t of demersals and 1.5 million t of pelagics (Table 10.13). The total standing biomass was estimated at nearly 1.6 million t at a time that landings were at a fairly modest level, probably below 200,000 t (Table 10.14). Both estimates indicate that there should be a potential for an expansion of fisheries.

Table 10.13 Eastern Central Pacific: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

The survey programme of the DR. FRIDTJOF NANSEN consisted of four surveys of the shelf from Colombia to the Gulf of Tehuantepec in different seasons during 1987.

The main feature regarding fishery oceanography was the confirmation of the disruption of the stable tropical conditions of the surface layers by the well-known seasonal upwellings in the Gulfs of Panama, Papagayo and Tehuantepec caused by strong winds blowing through passages in the mountain ranges between the Atlantic and the Pacific.

In addition to inshore assemblages of small pelagic fish dominated by thread herring and carangids and demersals (butterfishes, grunts and snappers being the most common), there were offshore assemblages in deeper waters in the highly productive areas consisting of sea basses, silver smelts and hairtails. Deep-water shrimps and langostino (Pleuroncodes planiceps) were found in high abundance off Nicaragua and El Salvador and squids in the Gulf of Panama. Giant squid occurred off the shelf.

Table 10.14 Eastern Central Pacific: Summary of total biomass, densities and recent reported annual landings. (Landings for Panama represent total for country)

An overview of the estimates of total biomass of standing stocks, densities and recent landings is shown in Table 10.17. The effects of processes of enrichment of the surface layers by the seasonal upwellings are clearly evident in the Gulf of Panama and in Nicaragua from the Gulf of Papagayo upwelling. Guatemala has joint stocks with southern Mexico and may be affected by the upwelling there. The biomass density on the northernmost shelf in the Gulf of Tehuantepec was not particularly high, but the high densities of mesopelagic fish found off the shelf are not included in the biomass estimate. Their abundance may relate to surface waters enriched by upwelling and advected offshore by the strong winds in the Gulf.

The landings are low compared with the biomass levels. It may not be commercially feasible to utilize fully all small pelagic fish, but there is no doubt significant potential for expanding fisheries on many stocks, particularly in Nicaragua, Panama and Guatemala.

10.3.7 Western Central Atlantic

The DR. FRIDTJOF NANSEN surveys covered only a small part of this statistical area, viz. the north coast of South America from Suriname to Colombia. This includes an upwelling area off the eastern part of Venezuela, that was already known in the 1960s. Despite the incorporation of this area in the estimate of potential for the Caribbean a a whole, Gulland's (1970) potentials of demersals (125,000 t) and small pelagics (600,000 t) are low (Table 10.15). The estimate for the area between Trinidad and Brazil was 200,000 t for demersals and 200,000 t pelagics.

The DR. FRIDTJOF NANSEN estimates of the total standing biomass for Suriname + Guyana amounted to 950,000 t, indicating a somewhat higher potential (Table 10.16). The biomass of pelagics off Venezuela Oriente is very high and compared with Gulland's estimate of pelagics for the whole Caribbean. The landings from this resource appear to be reasonably high.

Table 10.15 Western Central Atlantic: Summary of estimates of potential catches of demersal fish and small pelagic fish extracted from Gulland (1970)

Four surveys of the shelf from Suriname to Colombia were undertaken in different seasons during 1988.

The main known features of fishery oceanography were confirmed by survey observations. Sloping isopycnals set up by the current off the Suriname-Guyana coast do not demonstrate true upwelling, but deeper water is brought onto the shelf and some enrichment of surface layers occurs through vertical mixing. Discharges of large amounts of freshwater affect wide parts of the surface water in this region. In Venezuela there is intensive seasonal upwelling along the eastern coast, but stable surface layers off the west coast. In Colombia there is modest seasonal upwelling off the Guajira Peninsula, but elsewhere surface layers are rather stable.

Table 10.16 shows a review of the mean total biomass of the standing stocks, densities and recent landings by regions. The effects of the processes of enrichment of the surface layers on standing stock biomass are evident in Suriname and the eastern Venezuelan coast of the Caribbean.

Table 10.16 North coast of South America: Summary of total biomass, density and recent reported annual landings (Landings for Venezuela represent total for country).

Statistics of reported landings in 1988 and 1993 show generally low levels compared with the biomass. It is not evident, however, that there may be a commercial basis for utilizing fully the small pelagic fish in for instance Suriname, but there is no doubt a considerable potential for growth of fisheries in this country. There has been a recent substantial increase in the landings in Venezuela, but there may still be potential for further growth.

10.4 CONCLUDING REMARKS

It is hoped that the review presented in this book has provided the reader with an appreciation of how the survey programme with the DR. FRIDTJOF NANSEN has contributed to increasing knowledge on the availability of fishery resources of the ocean in the last few decades.

In a rapidly changing environment, as a consequence of anthropogenic activities, the data accumulated by the old DR. FRIDTJOF NANSEN, and still continuously expanded by surveys with the new vessel, represent important historical evidence on the composition and abundance of fishery resources in shelf areas of the world. These data may be very valuable for further studies of the marine fisheries resources of several productive areas and further work should be encouraged.

During the editing of this document the book “Ecological Geography of the Sea” by Alan Longhurst was published in 1998. Longhurst (1998) provides an overview of biological productivity in all areas. The results of the surveys of the DR. FRIDTJOF NANSEN provide an interesting fisheries complement to Longhurst's review.