3 SURVEYS IN THE ARABIAN SEA AND ADJACENT GULFS

3.1 RESEARCH HISTORY OF THE INDIAN OCEAN AND OVERVIEW OF ASSIGNMENTS OF THE DR. FRIDTJOF NANSEN

Until the 1950s, the Indian Ocean had to a large extent been neglected by marine scientists compared with the long history of explorations and investigations of the Pacific and Atlantic Oceans. This was one of the reasons why, following successful international co-operation in oceanographic research in the Pacific and Atlantic Oceans during the International Geophysical Year (IGY) 1957, oceanographers agreed to collaborate and concentrate oceanic investigations in the Indian Ocean for a period of several years starting in 1959. This evolved into an international venture in marine science on an unprecedented scale: the International Indian Ocean Expedition (IIOE), which during 1959/65 attracted about 40 research vessels from 20 countries at a total estimated cost of about US$ 60 million (Behrmen, 1981).

The Indian Ocean had the attraction of the unknown, but another main reason why marine scientists from so many countries joined the IIOE was the professional interest in this unique area, closed by land masses in the north in contrast to the Pacific and Atlantic Oceans, and with its regular seasonal reversal of circulation caused by the monsoon shifts.

Although not all expectations for a systematic coverage of observations in time and space were met - the venture has been described as a five-year series of more or less unco-ordinated expeditions - the IIOE resulted in a greatly increased state of knowledge of all basic processes in the Indian Ocean. The fishery sector had been held out as the main potential beneficiary of the IIOE, but direct fisheries research was not among the types of investigation made. Nevertheless, the data and information on physical, chemical and biological oceanography acquired were of great interest for fisheries.

FAO's Fisheries Department was keenly awaiting the presentation of the findings of the IIOE and commissioned a special analysis of data relevant to fisheries which was submitted in the form of an “Atlas of the Arabian Sea for Fishery Oceanography” (Wooster et al., 1967). Reviewing data on primary production and considering the fishery implications, the authors of the Atlas concluded: “that it may quite confidently be stated that the western side of the Arabian Sea is one of the more productive parts of the world oceans”.

The mean productivity of the entire western portion of the Indian Ocean up to several hundred miles offshore was reported to be on average several times that of the mean of the world oceans with an estimate as large as or larger than that encountered in such upwelling areas as the eastern boundary currents off Peru and off West Africa. Since the Peruvian fishery in its contemporary remarkable expansion had by 1967 grown to give an annual yield of 8 to 9 million t, the comparison with its primary productivity was a very optimistic signal for Indian Ocean fisheries. One of the authors of the atlas, Schaefer, estimated later a potential annual yield of the Arabian Sea of 10 million t (Shomura et al., 1967) and fishery adviser Wib Chapman, in one of his widely circulated discussion letters at that time indicated a range of 10–20 million t.

The Indian Ocean had been a largely neglected area not only with regard to scientific research, but also fisheries development. Following the encouraging findings from the IIOE of high primary productivity, FAO took the lead in creating a programme for support and co-ordination of fisheries research and development in the Indian Ocean. After a preliminary discussion at the First Session of the FAO Committee on Fisheries (COFI) in 1967, at its Second Session in 1968 COFI recommended the establishment of an Indian Ocean Fishery Commission (IOFC) having as one of its major objectives the promotion of research and development activities in the area through international efforts and in particular with the assistance of international aid programmes. A special International Indian Ocean Fishery Survey and Development Programme or Indian Ocean Programme (IOP) was planned at IOFC's First Session in 1968 and was subsequently established with financial support from UNDP.

In its first phase, estimated to last one year, the IOP was to assemble and examine all existing data and information so as to provide a synopsis of the current status of knowledge regarding the resources of the area. This was to be followed by an operational phase of three years and a concluding one-year period.

The final operational plan of the programme, presented in Marr et al. (1971), included a synopsis of special consultations and detailed studies made during its preparatory phase. This plan first reviewed all the important questions regarding resources and potential yields. A simple catch projection for the Indian Ocean as a whole was made by extrapolating the catch per unit area obtained in the Pacific and Atlantic Oceans based on the not unreasonable assumptions that these catch rates should be roughly comparable. In 1968 the actual catch per unit area in the Indian Ocean was only about one-sixth of that of the other two oceans and about one fourth if the areas compared were those of the continental shelves. Since the Pacific and Atlantic Oceans were at that time assumed to produce only about half of their maximum yields, the total potential of the Indian Ocean would on this comparative basis be some 8 or 12 times the 1968 catch.

A more comprehensive analysis of Indian Ocean potential, based both on observations of productivity and on extrapolations and comparisons of yield or biomass densities presented by Shomura (in Gulland, 1970) estimated the total yield to be about 14 million t, or some six times the 1968 catch. This level was used in the operational plan of the IOP to demonstrate possible growth rates of Indian Ocean fisheries over a 20-year period.

The review by Marr et al. (1971) also made reference to other available studies of the Indian Ocean's potential. Based on observations of primary productivity and considerations of trophic relationships, Cushing (1971a and 1971b) estimated annual production at the third trophic level at about 9.5 million t for the upwelling areas, and Prasad et al. (1970) gave an estimate based on ratios of carbon production and results from exploratory surveys of 11 million t annual sustainable yield. Although these estimates indicated a lower total potential yield than 14 million t, they agreed in holding out considerable scope for expansion of the fisheries.

Among its various recommendations, the IOP's Plan for Fishery Development in the Indian Ocean contained a series of pre-investment fishery development surveys. Priority was given to a pelagic fish assessment survey in the northwest Arabian Sea to be conducted by one vessel over one year. This part of the ocean had been reported in several studies as one of the most promising and it was thought that the high productivity would sustain abundant stocks of small pelagic fish as in the Pacific and Atlantic Oceans in upwelling systems of eastern boundary currents.

Survey assignments

The planning of the IOP preceded that of the DR. FRIDTJOF NANSEN programme by about one year. FAO and UNDP saw the little-researched Indian Ocean as an important potential region of operation for the DR. FRIDTJOF NANSEN and the high-priority task of an acoustic-cum-fishing survey of the small pelagic fish in the northwest Arabian Sea represented a very suitable first assignment. This initiated a long period of investigations with the vessel in this region. Nine of the first ten years of operation (1975–84) were spent in various parts of the Indian Ocean (Table 1.1 and Appendix II). This reflected the special need which existed for more information on its fishery resources and also demonstrated the expectations as to its development potentials.

Figure 3.1 shows the geographical coverage of DR. FRIDTJOF NANSEN surveys in the Indian Ocean: most of the Indian Ocean's coastal areas were included, except those of India, Australia and most of Indonesia. The southwest coast of India was already covered by a survey programme with similar objectives, the FAO/UNDP Pelagic Fishery Project (IND/69/593) in the period 1971–76 and its successor (Pelagic Fishery Investigations on the Southwest Coast - Phase II (IND/75/038) (FAO, 1982).

Figure 3.1 Location of the DR. FRIDTJOF NANSEN surveys in the Indian Ocean and South China Sea, 1975–84

During the first two years (1975 and 1976), the DR. FRIDTJOF NANSEN surveys differed from later assignments in having an exploratory character, investigating wide and largely unknown areas in order to obtain a first appreciation of the distribution, composition and the magnitude of the pelagic fish stocks in these waters. In retrospect, this first exploratory phase could be considered unnecessarily long, and a change to more detailed investigations of specific areas could have been made after only one year. The situation after the completion of one pre-monsoon and one post-monsoon coverage was, however, one of considerable uncertainty (IMR, 1976b). Although survey results confirmed the occurrence of small pelagic fish in the known highly productive inshore areas, their estimated abundance was nowhere as high as expected. On the other hand, the very high abundance of mesopelagic fish over the whole survey area was an unexpected finding. There was thus a need to confirm these general results and check on possible inter-annual variations. The character of the work and main objectives were therefore maintained in the continued survey, although there was some redistribution of survey intensity with more attention being given to the most promising parts.

Subsequent assignments were based on a different approach as regards both the general organization of the surveys and their objectives. Even though still operating under the umbrella of the IOP until its termination in 1979, each assignment was now planned and executed in closer co-operation with authorities of the countries concerned and the survey period was estimated to allow detailed repeated investigations of all the resources which could be targeted by the methods used as well by environmental studies.

In the late 1970s there was considerable interest in fishery research among the coastal countries of the Indian Ocean. FAO, through IOFC and IOP, had made the countries aware of the potentials for fishery development. In addition most States in the region had by the late 1970s established EEZs in accordance with the provisionally agreed text of the Law of the Sea Convention and were conscious of a need for more information on the fishery resources within their EEZs.

The sequence of new DR. FRIDTJOF NANSEN assignments did not follow a long-term plan, but was adjusted to meet priorities set in part by FAO/UNDP, and in part by NORAD. In many cases assignments were renewed in a region or coastal zone already covered. The objective was then to confirm and supplement previous work and to study interannual fluctuations of the composition, distribution and abundance of the resources.

The IOP's original plan for a pelagic fish assessment survey of the North Arabian Sea included the whole shelf and adjacent ocean from Somalia to Cape Comorin (the southern tip of India) (Midttun et al., 1973). In order to provide a more complete overview of the pelagic resources of the entire North Arabian Sea reference will also be made to the findings from the almost contemporary (1971–75) survey programme off the southwest coast of India, the FAO/UNDP Pelagic Fishery Project (IND/69/593). This project was not part of the DR. FRIDTJOF NANSEN programme, but IMR was involved in its scientific execution and there was an important intercalibration between the project vessels RASTRELLIGER and DR. FRIDTJOF NANSEN. In order to maintain a time sequence in the review, this project will be presented first (see Section 3.2).

Section 3.3 deals with the first exploratory period of the DR. FRIDTJOF NANSEN which covered the highly productive northwest Arabian Sea from Pakistan to Somalia in 1975–76. Relevant findings of the Pakistan assignment of January-June 1977 are also included.

The joint findings of these surveys represented at that time the first, and in retrospect apparently fairly conclusive, replies to the important questions concerning the fish potentials of the Arabian Sea and adjacent Gulfs for which such high expectations had been held out.

The mesopelagic fish, which in their high abundance are restricted to the slope of the continental shelf and the adjacent oceanic parts of the northwest Arabian Sea, are described separately in a section which includes the special follow-up surveys mounted for these species in 1979, 1981 and 1983 (Section 3.4).

Section 3.5 describes the follow-up surveys for small pelagic and demersal fish from Pakistan to Somalia from 1983 to 1984 on a country-by-country basis.

Other parts of the Indian Ocean are dealt with in Chapters 4 and 5.

3.2 PELAGIC FISHERY INVESTIGATIONS OFF SOUTHWEST INDIA, 1971–75: RESULTS OF THE FAO/UNDP PROJECT IND/69/593

Project objectives and effort

The southwest coast of India (Malabar coast) is included in this review for reasons of completeness. This will allow comparisons between the upwelling system off the Somalia-Arabian coast and that off the Malabar coast. The survey methods used were more or less identical to those of the DR. FRIDTJOF NANSEN programme, a result of IMR's involvement in both programmes.

The project resulted from a request from the Government of India to UNDP/FAO in 1967. The background was the experience of wide fluctuations in the yields of the important inshore fisheries for oil sardine (Sardinella longiceps) and mackerel (Rastrelliger kanagurta) on the coast from Cochin to Goa resulting in shifts between seasons of glut and years of failing fisheries with extremely low landings. It was envisaged that an expansion to offshore fishing could stabilise production, but prior to such attempts investigations were required to determine the distribution, migrations and abundance of the two species. Such investigations were the main objectives of the project with the addition of fish behaviour studies related to fishing methods as well as experimental fishing.

Equipment and installations of importance for the project were available at the headquarters of the former Indo-Norwegian project at Cochin. This institution continued as a Government of India organization after the termination of the Norwegian co-operation and was given responsibility for the project on the Indian side. The major FAO/UNDP components were sub-contracted to the Norwegian Agency for International Development (NORAD) with the IMR acting in the field.

The main project work was to be based on acoustic-cum-fishing surveys with two research vessels. Project duration was about five years, 1971–75. Following a later decision, a second phase (IND/75/038) with FAO as the executing agency, was continued until 1979. The long-term objective of this second phase was to assist in developing the pelagic fishery off the southwest coast within the framework of government policy as expressed in the five-year plan 1974–79. Principal elements of that phase were vessel and gear development, fish handling, transport, utilization and marketing. Although resource surveys were mounted, their results contributed no information additional to that provided by the first phase and thus these surveys are not included in this review.

The effort expended in the first phase was considerable, with two project research vessels, RASTRELLIGER (46 m LOA) for offshore and SARDINELLA (16.5 m LOA) for inshore waters, support of exploratory vessels from the former Indo-Norwegian project and additional aerial surveys for supporting observations of surface schools. Personnel in the first phase totalled 35 man-years of expatriate and 55 man-years of national staff. Training through fellowships and on-the-job was an important element of the project.

Environmental components of the project included oceanographic investigations with particular reference to the effects of monsoon shifts and related upwelling process, and zooplankton and fish egg and larva investigations to study production and spawning cycles.

The work programme was conducted successfully, with dense coverages both in time and space by the survey vessels of the shelf between the Gulf of Mannar in the south and Ratnagiri in the north. A very large amount of data and material pertaining to all the objectives of the project was collected in a systematic way, processed and reported on. This was a unique and major research programme, an intensive study of an assemblage of important pelagic resources affected by distinct seasonal environmental changes in a highly productive tropical upwelling area. The results are described in detail in 18 Progress Reports (see references under IMR/NORAD/FAO), and a brief summary is available in a Terminal Report, (IMR/NORAD/FAO, 1976h). Participating Indian scientists have reported special scientific findings in Indian journals, but a historical report with a comprehensive description and analysis of the results of the pelagic fishery investigations on the southwest coast of India has unfortunately not been attempted.

The project is of interest also regarding the investigation methods used. The preceding Indo-Norwegian project initiated sea-going research already in the late 1950s, but this consisted mainly of exploratory and experimental fishing with some oceanographical work. The emphasis in this new project was placed on systematic acoustic-cum-fishing surveys, using an echo integrator. The acoustic integration technique was, however, soon found to be inadequate in surveying the predominantly surface-schooling oil sardine and mackerel and therefore a multi-vessel school survey with synoptic observations from aerial surveying was tried. Echo integration was applied for assessing other pelagic and benthopelagic stocks. In retrospect it is seen that in this initial phase of the echo integration technique, problems of its application were unavoidable. The most serious constraint was the lack of a reliable method of instrument calibration (see Section 2.3).

Figure 3.2 shows the project area and examples of coverage during the oil sardine/mackerel and echo integration surveys. The shelf width to 200 m depth increases from about 30 nmi in the south to about 60 nmi in the north. Data on the survey area are given in Table 3.1 which also shows approximate estimates of the resulting survey intensity of the course tracks shown in Figure 3.2. The estimate is based on the area covered for the oil sardine/mackerel survey and on the total shelf area for the general surveys. The survey intensities are high compared with efforts in similar exercises.

Figure 3.2 Example of course tracks, hydrographic stations and fishing trials in a survey coverage of the project area (from IMR/NORAD/FAO, 1976g)

Table 3.1 Survey area and approximate degree of coverage

The project's main survey activities are summarized in Table 3.2. The SARDINELLA was available from June 1971 and the RASTRELLIGER from January 1973. The post-monsoon season, September-October, was selected for synoptic surveys of the schooling oil sardine and mackerel while the numerous main integration-cum-fishing surveys covered all seasons, with the RASTRELLIGER also during the southwest monsoon. There are thus four post-monsoon surveys of the oil sardine and mackerel schools, 1972 through 1975, and three years of echo integration surveys, 1972–73, 1973–74 and 1974–75.

Table 3.2 Summary record of the project's main resource surveys 1971–75 using the research vessels SARDINELLA and RASTRELLIGER, auxiliary vessels and aerial scouting

A programme of oceanographic observations in seven fixed sections was followed from June 1971 to October 1975 with observations of temperature, salinity and dissolved oxygen. No comprehensive reporting of the project's results as regards experimental fishing is available, but findings are discussed in the progress reports.

Hydrography

The profound seasonal changes in the environment represent keys to the understanding of many features of the distribution and behaviour of the fish stocks. The oceanography programme is described in IMR/NORAD/FAO 1973b and 1976e and in Johannessen et al. (1987). During the northeast monsoon, November-March, there is a weak northward-flowing current along the coast transporting low salinity water originating from the Bay of Bengal. The surface layer is stable and the water on the shelf is well aerated. The current reverses in March-April responding to the main wind field in the northwest Indian Ocean, and during the southwest monsoon season from May-June to September-October the current flows southward at a much higher velocity than the northward current.

In the southward current the isopleths tilt upwards towards the shore. This appears to start already in March-April. With the start of the southwest monsoon in May-June the sloping intensifies and upwelling occurs along the coast with increased primary production. There has been some uncertainty regarding the forces driving this upwelling, but according to Longhurst and Wooster (1990) the prevailing westerly local monsoon has an equatorward component which advects surface water offshore leading to the upwelling of water already tilted by the anticyclonic gyre spun by the southwest monsoon.

Figure 3.3 A: Depth of the intersection between the 1 ml/1 oxygen isoline and the bottom
B: Zooplankton biomass in ml plankton per m³ water filtered.
1: Karwar section; 2: Kasaragod section; 3: Cochin section; 4: Quilon section from Johannessen et al., 1987

The shoreward tilting of the isopleths in the southward current and the upwelling causes penetration of low-oxygen water, less than 0.5 ml/l, over the entire shelf almost on to the shore. The regularity of this process and hence of the upwelling is demonstrated in Figure 3.3 which shows the intersection between the 1 ml/l oxygen isoline and the bottom in four sections along the coast from 1971 to 1975. The direct effect of upwelling on production is demonstrated by the observed densities of zooplankton biomass in these sections. It is notable that the minimum zooplankton biomass is observed in February-March with the trend of increase starting already in April-May prior to the onset of the local southwest monsoon.

The upwelled low-oxygen water affects the fish distribution on the shelf. The surface layer of aerated water which forms the habitat of the oil sardine and the mackerel is very thin and restricts the depth distribution of these fish. Changes in the distribution along-shore were also noted, both for demersal fish and anchovies with restriction to the southern shelf and Gulf of Mannar in the southwest monsoon period.

Biomass estimates from the echo integration surveys

The reliability of the abundance estimates using this method depends inter alia on whether the constants used for converting acoustic indices to biomass were appropriate for the various types of fish and based on a proper calibration of the system. As discussed in Chapter 2, the methods used in the first part of the 1970s of measuring the performance of the echo integration system had a low precision. The hydrophones were generally unreliable and their use under field conditions such as in the project reviewed here were especially difficult. Under these circumstances, experimental determination of the conversion factors by cage calibration on live fish as chosen by the project was the best approach, but this involved cumbersome and time-consuming experiments not suitable for regular monitoring of the instruments.

The performance of the 120 kHz EKS sounder which was used throughout for echo integration was reported in IMR/NORAD/FAO (1974a) to be 104.9 dB (probably the level measured after installation in Norway). Cage calibrations with catfish (Arius thalassinus) showed a mean target strength of -29 dB/kg corresponding to a conversion factor C of 6.5 t/nmi²/mm. Further experiments were made in 1974 (Nakken, 1974), which confirmed this value of C for 17 cm fish and indicated a stable performance of the system. However, -29 dB/kg seems a high target strength level for these fish. A more likely level is -32 dB/kg which would correspond to an instrument performance of 108 dB.

Unfortunately, cage calibrations with anchovies were not possible. Although an attempt was made to determine the target strength level of these small-sized fish (IMR/NORAD/FAO, 1974a), the proportional relationship between fish size and conversion factor C which solves this problem was only adopted by IMR after the termination of the project. Assuming a size of the anchovy of 8 cm for the seasons of their high abundance, the conversion factor C for estimating biomass would have to be reduced by about 50% to adjust for this size of fish.

The instruments of the research vessels were maintained and kept operational by the project's experienced electronic engineers. The maintenance reports are no longer available, but through interviews it has been established that a major repair with replacement of electronic parts was undertaken in late 1974. No further cage calibrations were made, but in June 1975 an intercalibration was performed with DR. FRIDTJOF NANSEN which showed that

120 kHz RASTRELLIGER = 120 kHz DR. FRIDTJOF NANSEN * 7.9

The performance of the 120 kHz system of the DR. FRIDTJOF NANSEN was later found to be 103 dB in calibrations with a metal sphere (standard target). If this was the performance also in 1975, the 120 kHz echosounder of the RASTRELLIGER must have had a performance of 112 dB. This indicates an increase in performance after the repair of 112 - 108 = 4 dB or 2.5 times.

The details of this assumed history of the performance of RASTRELLIGER's system are not much more than guesswork, but the intercalibration strongly suggests that a significant increase of the performance must have taken place between 1974 and 1975.

The project's time-series of the biomass estimates of anchovies shows an approximate doubling from the 1974 to the 1975 season. Substantial increases of biomass of the order of three to seven times from 1974 to 1975 were also reported for the other groups covered by the echo integration surveys. It seems difficult to find a biological explanation for such large and simultaneous increases in the abundance of all the groups of both short-lived and long-lived fish.

The survey programmes included many fishing trials with the pelagic trawl for identification and biological sampling. Although a simple relationship between catch rates and abundance in these trials cannot be expected, an overall biomass increase of about three times for the target fish should be reflected in the catch rates. Table 3.3 shows mean catch rates in comparable series of data from the fishing programmes of the two vessels RASTRELLIGER and SARDINELLA taken from the Cruise Reports (IMR/NORAD/FAO, 1974/75).

There is no reflection in these data of generally higher catch rates in 1975 than in 1974. The July-August 1975 data are from fishing operations on concentrated anchovy in the Gulf of Mannar, and are therefore not comparable with the other data.

This tends to confirm that the observed high abundance in 1975 resulted from an increase of the performance of the echo integrator system, most likely through a change in the TVG-receiver gain function.

Table 3.3 Comparison of catch rates in fishing trials for identification and sampling (kg/haul)

With small claims to accuracy, a revision of the biomass estimates has been made based on a reduction by 50% of the anchovies throughout due to their small size (see above) and a reduction of all 1975 estimates by 2.5 times corresponding to an increase of the integrator deflections from a 4 dB higher performance. The final results are presented in Table 3.4 and Figure 3.4.

Table 3.4 Revised biomass estimates from the echo integration surveys by target groups (1,000 t)

Figure 3.4 Adjusted estimated total fish biomass from echo integration surveys (for adjustments see text)

The maxima of the anchovy biomass are now seen to vary between 260,000 and 405,000 t and the total biomass of all groups covered by the echo integration surveys reaches levels of about 1.1–1.2 million t each year. The anchovies show the clearest seasonal fluctuations largely reflecting an annual production cycle in these short-lived fish. For the other groups the variation of the estimates between surveys is in addition to survey variance, likely to have been caused by seasonal changes in availability of the stocks which may have a distributional area which exceeds that covered by the surveys.

This comprehensive and long-time series of biomass estimates shows features which give some grounds for confidence. Apart from the obvious seasonal cycle for anchovies, there is a degree of consistency in the proportion between the groups which tends to confirm their proper identification, and there is some regularity in the seasonal variation in total biomass. Although allowance must be made for underestimation due to bias caused by signal saturation effects in the scientific echosounder/integrator system of the first generation, the estimates are still of historical interest as they represent a stage of low exploitation of the stocks.

Experimental fishing

A planned report consolidating the project's work on experimental fishing did not become available, but several progress reports refer to this. Thus in reporting on the first six months work in 1971 with SARDINELLA (IMR/NORAD/FAO, 1971) the experience of purse-seine trials on adult mackerel is said to confirm parallel results obtained earlier by the Indo-Norwegian project's M-boats that these schools of mackerel are readily caught with suitable purse-seining techniques.

In describing the offshore distribution of oil sardine and mackerel demonstrated by the 1971/72 surveys (IMR/NORAD/FAO, 1973a) the possibility of extending the fishing season through offshore purse-seining is indicated. Reference is again made to successful purse-seine trials of the Indo-Norwegian Project with various sizes of boats and purse-seines. The point is made that an offshore fishery would increase the exploitation of large adult fish which are present in the inshore fishery only to a small extent. The offshore fishery would thus only partly exploit the resources on which the traditional fishery depends. This point is also made in IMR/NORAD/FAO (1974a), where reference is made to purse-seine catches of oil sardine of up to 40 t/set by the SARDINELLA with her 180 fathom seine as well as to the successful fishing of offshore schools by various vessels of the Integrated Fisheries Project. It is concluded that the technique of an offshore purse-seine fishery is well established, but it is recommended that before attempting to develop such a fishery a pilot-scale programme of experimental and demonstration fishing be conducted.

Mid-water trawls were tested on anchovies and other types of fish. In IMR/FAO/NORAD (1974a) night fishing with SARDINELLA's 100 m² trawl was reported to yield catches of 1.2 t/h. In IMR/NORAD/FAO (1976a) catch rates of anchovies with RASTRELLIGER's 400 m² trawl in the Gulf of Mannar are reported to have been about 4 t/h during daytime and about 13 t/h at night. Reference is also made to purse-seine sets of 1–4 t of scads by the Integrated Fisheries Project's vessel SAMUDRADEVI. Catch rates of “other fish” of 1.6 t/h were reported for SARDINELLA's small pelagic trawl. In IMR/NORAD/FAO (1976b) mackerel and oil sardine purse-seine sets of 0.4 to 5.5 t are reported for SARDINELLA and of 2 to 27 t for SAMUDRADEVI. Up to 1.6 t of catfish and ribbonfish were obtained with the mid-water trawl and up to 1 t/h of shallow-water mix and 3.8 t/h of “other fish”.

Through project activities in this field of fishing experiments as well as that of the previous Indo-Norwegian Project and the Integrated Fisheries Project the feasibility of offshore purse-seining for mackerel and oil sardine and perhaps also scads seemed to have been well established. It was also demonstrated that “interesting catch rates” of anchovies, shallow-water mix, ribbonfish and catfish and “other fish” could be obtained by mid-water trawl.

Oil sardine and Indian mackerel

The assessment of the stocks of oil sardine (Sardinella longiceps) and Indian mackerel (Rastrelliger kanagurta) was a main objective of the project and the method to be used was carefully considered. The first period of exploratory surveys with SARDINELLA showed that in the post-monsoon months August-October the two species could be found in belts of surface schools 6–20 nmi offshore extending for at least 200 nmi along the coast (IMR/NORAD/FAO, 1973a). Further observations indicated that the oil sardine and the Indian mackerel mainly occurred as dense schools and that school formation was maintained also during the night even if they were then less densely packed. With this near-surface distribution it was concluded that the echo integration technique could not serve as the main method of biomass assessment because of avoidance of the vessel and it was decided to try aerial surveys combined with visual, sonar and echosounder observations from vessels. The best period for the aerial survey was judged to be September-October when the schools appeared to have an extreme surface-bound distribution prior to their inshore migration.

Surface schooling is common behaviour among small pelagics, especially sardinellas and sardines, but usually interspersed with occurrence in schools or layers at greater depths. The more consistent restriction to the surface layer of the two species along the Malabar coast may be related to the limits set by the low oxygen content of the water at greater depth. When school-counting with sonar was adopted as a part of the routine survey techniques in 1974–75, observations demonstrated that sardine and mackerel occurred in schools on the shelf through most of the year also in the southwest monsoon period when, due to rough weather, they could not be seen at the surface (IMR/NORAD/FAO, 1976d).

Table 3.5 Summary review of results of oil-sardine and mackerel post-monsoon aerial and vessel surveys (totals and separated by species where possible)

The main data from the series of annual oil sardine and mackerel surveys are reviewed in Table 3.5. The basic data are the estimated number of schools, their mean volume and packing density.

In the 1972 survey the main effort was by aircraft and only one vessel, the SARDINELLA, provided supporting data on submerged schools that had been missed by the aerial survey and additional observations for estimation of schools dimensions and packing densities. The project assessed that this survey had produced an underestimate and it was not included in later summaries. The calculation of school volume was, as in 1972, based on echosounder observations of submerged schools. This may have introduced a bias as effective vessel avoidance by schools may be dependent on school size. Estimates of the size of sardine schools from aerial photographs was on average only a quarter of that from echosounders.

The basis for the 1973 survey was a much larger combined effort by aircraft and vessels which would be expected to give an improved estimate of school numbers. In the 1974 and 1975 surveys, school volumes were estimated from sonar records where vessel avoidance is not critical, and these showed volumes about 10 times smaller. The aerial observations in 1974 were of more limited value partly because of unsuitable weather, and also because fish behaviour differed from previous surveys with more schools being submerged. A systematic sonar coverage was, however, made of the whole area and there would not seem to be any apparent bias in the estimate of total school numbers. The last survey, that of 1975, was without aerial support and consisted of a multi-vessel effort based on sonar records and visual surface observations. In this series of surveys from 1973 to 1975 the data on total school numbers is likely to be the most consistent, perhaps not in terms of precision, but as regards year-to-year changes in serious bias.

In 1973 and 1974 separate assessments were made of the sardine and mackerel stocks based on aerial identification. In 1975 the assessment was combined for the two species, but based on samples of observations of school identification, oil sardine was estimated to represent about 70% of the total.

An interesting point of methodology in this type of investigation is that from 1974 a correction was introduced for the effect of the elliptical shape of the schools. The previous assumption of circular shape had caused a negative bias.

Use of the sonar of the RASTRELLIGER transmitting at 90° to the vessel's course represented a new and it seems improved approach to the difficult problem of describing the distribution of the oil sardine and mackerel stocks and assessing their abundance (Smith, 1971; IMR/NORAD/FAO, 1975c). This method was applied in a systematic way from September-October 1974 onwards on all main surveys and the results were presented in the form of distribution charts and as estimates of the total areas within which schools were numerous.

The series of distribution charts based on the sonar method which cover one year from November 1974 to September-October 1975, shown in Figure 3.5, are among the project's principal achievements. These represent an adequate response to one of the project's main objectives: to describe the temporal and spatial distribution of these main species. When not presumably inshore as in December 1974 the stocks were found in well-defined school areas which cover large parts of the shelf including the Gulf of Mannar in the period February-March to July-August 1975. These stocks would thus be available to offshore fishing over a large part of the year.

Figure 3.5 Distribution of mackerel-and oil sardine schools from sonar surveys with the RASTRELLIGER, Oct 74-Nov 75. Source: IMR/NORAD/FAO, 1976b

Observations of school areas of oil sardine and mackerel were also part of the programme of the project's main fish surveys conducted by the two vessels SARDINELLA and RASTRELLIGER prior to September-October 1974, but they were then based on visual surface observations (IMR/NORAD/FAO, 1976a) and it seems doubtful whether it is appropriate to present the two sets of data as the time-series shown in the summary report (IMR/NORAD/FAO, 1976g). The extent of the distribution areas as determined from the sonar surveys only, shows an increasing trend from September 1974 onwards.

Whereas schooling behaviour, e.g., school size and packing density would be likely to vary with fish size, physiological state and environmental conditions, it is generally considered that it is largely unrelated to total stock abundance. There were as far as can be established no wide differences in fish size between the three post-monsoon surveys in 1973, 1974 and 1975 and they were conducted in the same season. It is thus not unreasonable to assume that the estimates of the total number of schools represent the best index of stock size. These numbers, 61,000, 101,000 and 255,000 for the respective years then describe rapidly increasing stocks from 1973 to 1975. This also conforms with the project recommendation that increased reliability can best be obtained by improved estimates of school size (IMR/NORAD/FAO, 1975c).

This trend may be compared to the landings in the inshore fishery (Table 3.6), where the totals for the calendar years from FAO's Yearbooks of Fishery Statistics Volumes 38,42 and 44 have been re-allocated with 64% and 36% in subsequent years as used by Longhurst and Wooster (1990) to estimate seasonal landings. There was a sequential decline in oil sardine landings from a level of about 200,000 t in 1970–71 to 92,000 t in 1973–74 followed by an increasing trend towards 1976–77 to about the previous high level which was maintained past the end of the decade. The landings must be expected to relate to stock abundance, but may also be affected by the inshore availability of the schools which may vary with seasonal variations in oceanography. The monsoon conditions as presented by Longhurst and Wooster (1990) do not show any special anomalies in these years and the trends of initial decline and subsequent increase over a six-year period indicates a stock effect. The project's time-series (1973–75) of total school numbers reflects this trend of increase although indicating a higher rate of growth. The estimates of total “school areas” from the sonar surveys also demonstrate an increase from the survey in September-October 1974 to that of October 1975.

Table 3.6 Landings of oil sardine and mackerel on the Malabar coast (1,000 t) (redistributed over subsequent years) and estimated total number of schools determined by surveys

One general conclusion of interest for survey methodology which may be drawn from the results shown in Table 3.5 and the discussion above is that observation of school volume and packing densities are the most critical parts of biomass assessment based on school surveys. Visual observations in the 1974 survey indicated that packing density could be calculated from an assumption of a distance between individual fish of 1.5 times the fish length. Some fishing experiments seemed to confirm this. Statistical data on the school volume estimates from the sonar surveys are not available. Estimation of school size by experienced skippers (IMR/NORAD/FAO, 1974c) showed a highly skewed distribution with 80% lying within a range of 2–10 t but reaching up to 60 t causing a low precision of the mean of 8. According to Marchal and Petitgas (1993), the sampling problem of increasing the precision of the school size estimate was the major impediment to improved biomass assessment of school surveys of small pelagics, mainly sardinellas in Venezuela. They found a similar strongly skewed distribution and suggested that school numbers which can be estimated with good precision may be a better index of stock size.

The mean school size of 4 t estimated for the 1975 survey may seem of the right order of magnitude. This average size indicates a total biomass of 1 million t in 1976–77. Judged from the levels of yield of the fishery this is likely to be an underestimate because it seems improbable that an inshore fishery conducted in a limited season only should take as much as 20% of the mean standing biomass.

Anchovies

The predominantly surface schooling behaviour of oil sardine and mackerel, the main target species of the project, was a setback for the researchers who had expected to be able to apply the echo integration technique in the investigations of these important stocks. In the introductory surveys with SARDINELLA in 1971–72 echo surveys had already revealed the presence of abundant resources of other types of fish on the Malabar shelf, and their investigation was included as a part of the work programme. Among these other fish, anchovies (whitebait) were found to be an important component of the pelagic community and the study of their biology, distribution and abundance was given considerable attention throughout the project. The behaviour of the anchovies made them a suitable object for the echo integration technique: they were mostly found in dense layers, sometimes dense schools at or near the bottom during daytime and in more dispersed layers higher up at night. The observations and material for these studies were derived from the main project surveys which covered the whole shelf with the multi-purpose objectives of acoustic recording, fish sampling, zoo- and ichthyoplankton studies, oceanography, and visual school mapping. A total of 25 such surveys were completed over the last three years of the project representing a very good seasonal coverage and a collection of an appreciable amount of information and material.

The anchovy stock was found to consist of several Stolephorus species of which S. heteroloba and S. bataviensis were the most abundant with S. zollingeri also occasionally abundant (Figure 3.6). The species were observed to be short-lived, reaching the adult size of 8–9 cm after about 6 months (IMR/NORAD/FAO, 1976g). Seasonal fluctuations in size composition and biomass would thus be expected and were indeed found. Anchovies were an easily identifiable acoustic target and the progress reports present distribution charts for each coverage which show a clear seasonal and quite remarkable pattern repeated in the data from 1974 and 1975. In January and up to April-May, the anchovies were distributed all along the coast. In the period of the southwest monsoon (July-August) and until November this group was almost exclusively found south of about 10°N with most of the biomass located in the Gulf of Mannar. Figure 3.7 shows the anchovy distribution in April-May and July-August 1975 respectively. This change of distribution was assumed to result from migration or southward displacement of the biomass with the fairly strong southward current set up by the southwest monsoon.

Figure 3.6 Relative abundance of anchovies (Stolephorus spp.) by areas for each coverage Oct-Nov 1974 to Aug-Sept 1975 (from IMR/NORAD/FAO, 1976b)

The production of the short-lived anchovy species must be expected to be closely related to that of zooplankton and perhaps to some degree also to the phytoplankton, both of which fluctuate sharply with the monsoon upwelling. The biomass estimates of anchovy by surveys (for adjustments see below) are shown in Figure 3.8. The zooplankton standing crop starts its seasonal increase in April-May, (Figure 3.3B and IMR/NORAD/FAO, 1974b) apparently prior to the start of the wind-driven upwelling. Nearly all the high anchovy estimates fall within the May-October period. The data indicate a 6–7 month production period, from March-April to September-October of the Stolephorus species with a southward displacement of the total biomass in July-August. This fits with the main trends in the observed seasonal appearance of larvae and juveniles.

Figure 3.7 Distribution and abundance of anchovies in April-May and July-August 1975, from IMR/NORAD/FAO, 1976b

Figure 3.8 Estimated anchovy biomass by surveys. For adjustments of original data in IMR/NORAD/FAO 1976g, see text.

It is possible that the survey area falls short of the total distributional area of the anchovies as indicated by the high biomass found in April-May 1975 in the northern subarea which may have originated further north. The project's data series on biomass show an apparent high variation between years. The highest estimates for biomass in the surveys of 1973, 1974 and 1975 were 520,000, 800,000 and 1,500,000 t respectively.

A review of these estimates and their basis, i.e., the state of the acoustic system and the target strength function was presented above.

Several other groups of fish were identified as recognizable targets of the echo integration system representing more or less co-occurring species. The annual summaries of the results of the main fish surveys presented in the progress reports include detailed descriptions of the distribution, composition and abundance of these groups as well as biological data on the most important species. The groupings used changed somewhat as experience was gained. Their general characteristics and species composition are described below (IMR/NORAD/FAO, 1976a and 1976b).

The scad group

The main species were torpedo scad (locally called horse mackerel) (Megalaspis cordyla), Indian scad (Decapterus russelli), trevallies (Caranx spp.) and less common frigate mackerels (Auxis spp.) and small tunny (Euthynnus affinis). These fish were mostly recorded as distinct schools in mid-water or close to the bottom, but also as single fish scattered over wider areas. They were most often found on the middle or outer shelf, but sometimes closer inshore or even beyond the shelf over deep-waters. Figure 3.9 shows examples of distribution charts for this group for May-June and July-August 1975. There is also a tendency for this group to have the highest abundance in the south during the southwest monsoon. The May-June distribution with the high abundance in the north indicates that the project area may only represent a part of the distributional area of this group and that changes in the abundance may be caused by migrations into and out of the area covered.

Figure 3.9 Distribution and abundance of the scad group in May-June and July-August 1975, from IMR/NORAD/FAO, 1976b

Figure 3.10 Distribution and abundance of “other fish” during the surveys in April-May and in August 1974, from IMR/NORAD/FAO, 1976a

Shallow-water mix

These fish, largely benthopelagic in behaviour, were grouped mainly because of their location within bottom depths of less than 15–20 m and because their distributions form close inshore belts. Typical species are ponyfishes (Leiognathus spp.), golden scad (Caranx calla) and butterfish (Lactarius lactarius). Especially the juvenile stages of a number of species from farther offshore were also found in this zone.

Other fish

This group did not represent a fish assemblage, but was used to describe the remaining part of the recorded biomass, although catfish (Tachysurus spp.) and ribbonfish (Trichiurus spp.), about 50% of this group, were often found together. These species are mainly benthopelagic and were treated as a separate group in the 1975 surveys. Their highest abundance was usually found on the mid-shelf at about 50 m of bottom depth and they were often the only species within relatively large areas. Other common species in the group were pomfret (Stromateus spp.), threadfin bream (Nemipterus japonicus), seerfish (Scomberomorous spp.), barracudas (Sphyraena spp.) and lizard fish (Saurida spp.). Bottom fish such as groupers (Epinephelus spp.), breams (Lethrinus spp.) and snappers (Lutjanus spp.) also contributed to the group as did sharks, rays and squids. Small amounts of mackerel and oil sardine found dispersed in the water column would also be included. In general the recordings of “other fish” were found over large parts of the shelf, usually scattered, but with some aggregated areas and with higher abundance in the south in the southwest monsoon season. Figure 3.10 shows examples of distribution charts of this group (including catfish and ribbonfish) for April-May and August 1974.

The proportions of the various groups of the total (adjusted standing stock) biomass estimates excluding schooling oil sardine and mackerel were as follows (mean over three years 1973, 1974 and 1975):

Summary of stock estimates

In the task of describing the composition, distribution and abundance of the surface schooling oil sardine and mackerel stocks, project scientists encountered problems of methodology, e.g., measurements of school sizes and packing density with which today's acoustic researchers are still struggling. Considerable experience was, however, gained over the period and the combined estimate of 1 million t for the 1975 season may well have been approximately right. It undoubtedly demonstrated the existence of a potential for expanding the exploitation of these stocks. The 1974–75 sonar surveys provided comprehensive information on the offshore distribution of these stocks throughout the year.

In retrospect it can also be seen that in this early phase of the echo integration technique there were considerable problems with its proper application and with full instrument control which must have affected the results. The rough revisions of these data have resulted in significant changes in some of the biomass estimates. The compilation of adjusted data shown in Table 3.6 may represent fairly realistic although probably underestimated biomass levels of the various resources.

Because of the seasonal changes in biomass and availability, the best estimate of standing stock size is likely to be closer to the maximum than the mean of each year's data. This approach would give stock biomasses from the echo integration surveys from Table 3.4 as follows (1,000 t):

Assuming an additional combined biomass of oil sardine and the Indian mackerel of 1 million t, the total biomass of neritic pelagic and benthopelagic fish in 1975 in the area from Palk Bay to Ratnagiri was thus in excess of 2.2 million t. This represents a mean density for the shelf area of 67 t/nmi² or about 200 kg/ha, a level found in highly productive systems.

It is of interest to compare this estimate of total standing stock with the annual yield of these types of fish in the area. Using data on landings by categories and states for 1989–1990 with the India Handbook on Fisheries Statistics (Ministry of Agriculture, New Delhi) and FAO's FISHSTAT Database as sources, an attempt has been made to estimate total annual catches in the project area. Almost all the reported catch of oil sardine and mackerel and roughly 60% of other pelagic and mesopelagic fish is according to the India Handbook from the project area and it is assumed that these proportions have remained unchanged since 1973. The additional groups comprise the following: Caranx spp., Carangidae, anchovies, hairtails, catfishes, Clupeidae and wolf-herrings. The group listed as “Marine Fish NEI” is also included as it is thought to contain a major component of small pelagic fish. The results are shown in Table 3.7.

Table 3.7 Approximate levels of landings of neritic pelagic and benthopelagic fish in the Project area (Malabar coast, southwest India) (1,000 t)

Mean annual landings for the mid 1970s (1973–1977) are about 330,000 t which compared to a standing stock level of 2.2 million t represents a very low fishing mortality and demonstrates considerable potential for expansion at that time. It seems that this has been and is being realized, first slowly with a mean catch of 420,000 t in 1983–1987, but approaching 700,000 t from 1989 on.

3.3 DR. FRIDTJOF NANSEN SURVEYS OF THE SHELF RESOURCES FROM SOMALIA TO PAKISTAN, 1975–77

The results of these surveys were briefly described in cruise reports (IMR, 1975; 1976a, 1976c, 1976d, 1977a), summarized in a Status Report (IMR, 1976b) and a Final Report (IMR, 1977b) and reported in a more comprehensive review by Kesteven et al. (1981).

Planning and execution of the surveys

A document setting out an elaborate plan to deal with all aspects of the survey work was submitted by IMR to the IOP in 1973 and this formed the basis of the survey plans (Midttun et al., 1973). The main target was the stocks of small pelagic fish, and the highest priority for survey areas was accorded to the coastal and upwelling zones from East Africa to the Gulf of Oman, including highly-productive offshore zones. It was uncertain, even doubtful if these productive offshore waters would hold abundant resources of small pelagic fish, but this represented an important issue to be solved. The two main seasons relating to the northeast and the southwest monsoons would need to be covered and a total survey period of two years was recommended. Weather conditions were not expected to affect work except at the height of the southwest monsoon. Acoustic observation and estimation was to be the principal method, supported by fishing with trawl and purse-seine for identification and sampling. The survey grid was to be flexible and adjusted according to fish density. Hydrographic profiles would be worked in selected positions for purposes of fishery oceanography. Plankton sampling for fish eggs and larvae would be made at hydrographic stations.

This plan was based on IMR's experience of acoustic surveys in the Norwegian and Barents Seas, but with additional experience on the part of some of its scientists from work in upwelling areas off Peru, West Africa and the southwest coast of India. Nevertheless, in scope and magnitude this was a unique undertaking not only for the IOP and IMR, but on a global basis. The task was to explore and investigate fishery resources off 3 410 nmi of coastline initially in a “blindfold” condition, since basic information on the behaviour and general distributional characteristics of the fish, normally used for deciding survey tactics and techniques, was not available from this largely unexploited ocean area. This first assignment of the DR. FRIDTJOF NANSEN was seen as an adventurous and exciting undertaking.

The region surveyed and its divisions are geographically described in Figure 3.11 and Table 3.8, adapted from Kesteven et al. (1981). The continental shelf is extremely narrow along the western shores of the Arabian Sea with mean widths ranging from 5 to about 20 nmi; this feature has important faunistic implications. The Pakistan shelf is wider with a mean of 30 nmi.

Figure 3.11 Sectors in first main surveys 1975–76

Table 3.8 Geographic specifications of the sectors of the surveyed region. Coastline length (nmi) and shelf area from 0 to 200 m (nmi²)

Table 3.9 shows the timing of the surveys in 1975–76 in relation to the southwest monsoon, and the survey effort in terms of distance steamed, inshore and offshore. The effort was reduced, especially in the offshore part after the first survey. The surveys covered the pre-and post-monsoon periods and also the northeast monsoon in the early part of the year. Sampling for identification was by pelagic and bottom trawls. Purse-seining was attempted, but the gear needed adjusting to local conditions of current and bottom which proved impracticable, and few trials were made.

Table 3.9 Timing of the surveys, surveyed distance (nmi) and number of fishing stations

Figure 3.12 shows the course tracks of the first survey and Table 3.10 presents operational records. The extent of the densely covered inshore waters is shown as surveyed inshore area, the distances surveyed are given for the inshore and the offshore areas, and the survey intensity defined as distance of observations relative to the square root of the surveyed space is estimated for the inshore area. This index is not available for the oceanic part of the survey where surveyed space is difficult to estimate. For the Gulf of Aden and Gulf of Oman estimates were made based on both the inshore course tracks on both sides and the crossings of the Gulfs. These gave mean survey intensities of 13 and 11 for the Gulf of Aden and Gulf of Oman respectively showing that they were as a whole well covered.

Table 3.10 Operational records of the main surveys: surveyed shelf area, surveyed distance and degree of coverage. Mean of 5 surveys

Figure 3.12 Course tracks of the first survey, February-June 1975

A comparison of Tables 3.8 and 3.10 shows that the estimates of the surveyed inshore space for all sectors exceeded the continental shelf areas thus demonstrating that the surveys extended well beyond the shelf. The degree of coverage was just about adequate.

Environmental observations included continuous recording of surface temperature and hydrographic profiles at selected positions as indicated by the course tracks in Figure 3.12. A total of 1,038 hydrographic stations were occupied in the five surveys, and 225 plankton samples were taken in the first two surveys. During the operations in 1976 seven current meters and temperature recorders were deployed and moored for six months on behalf of the University of Miami and the Institute of Marine Science in Kiel. The results of the oceanographic work carried out during the survey were analysed by Sandven (1979). They were essentially a confirmation of previous accounts and present no new features which could be directly related to the findings regarding the distribution and abundance of the resources. The environmental data are therefore not included in this review, but the data from the observations are available in the oceanographic databank system at IMR.

Small pelagic fish

The emphasis in this section is on the study relating to small pelagic fish which was the main objective of the surveys. Data on fish species recorded as “demersal”, but observed in mid-water and better classified as semi-demersal are included. Mesopelagic fish are reviewed in Section 3.4.

The small pelagic fish were only found inshore in close proximity to the coast. The charts of fish distribution of which Figure 3.13 is an example - presented for all the surveys as Figures 10 to 21 in Kesteven et al. (1981) - show these small pelagics to extend beyond the shelf in some areas especially off Somalia. A comparison with the course tracks, (Figures 2 to 6 in Kesteven et al., 1981) indicated that the seaward limits of the resources have generally been established. No records identified as small pelagic fish were made during any of the oceanic tracks beyond the slope.

In the first presentations of the results (IMR, 1976b and IMR, 1977b), the biomass was estimated from the integrator deflections averaged over 1 ° squares. In Kesteven et al. (1981) use was made of the plots of integrator deflections to describe the distribution patterns as shown in Figure 3.13 and to estimate biomass. The estimates from these two approaches showed no systematic differences by sectors and differed only about 10%, with a nearly identical overall mean of all surveys. Table 3.11 shows the biomass estimates by sectors and surveys as reported in Kesteven et al. (1981). For Pakistan, which was not included in that report, the estimates are from plots of the original integrator readings. All of these data should be analyzed as regards coverage, species identification and possible bias.

Table 3.11 Summary of original, unadjusted estimates of biomass (1,000 t) of small pelagic fish by sectors and surveys

Figure 3.13 Fish distribution from echo integration during the second survey August-November 1975

Identification of the sources of the integrator deflections presented a special problem in this largely unknown area with its diverse fauna. The expectation was a fauna of small pelagics, the most abundant of which would be sardinellas, especially the oil sardine, (Sardinella longiceps), mackerels (Scomber japonicus and Rastrelliger kanagurta), scads (Decapterus spp.), jacks (Caranx spp.) and anchovies (Stolephorus spp.).

Apart from Indian mackerel these species were indeed among the most common. But in parts of the area, especially the east coast of Somalia, the recordings of these fish of commercial interest were often mixed with those of non-commercial fish such as porcupine fish (Diodon sp.), splitfins (Synagrops sp.) and driftfish (Cubiceps sp.). More intensive sampling could have improved the identification, but this was at times prevented by unfavourable weather conditions and strong currents. In some cases it could only be concluded that the attempts had not been adequate and the data had to be rejected. One such case related to the coverage of the Gulf of Oman in Survey 5. A further review of the fishing trials revealed that the recordings off the east coast of Somalia in Survey 3 had not been properly identified as small pelagic fish and that they should be rejected. In general, sampling and identification on the Somali east coast proved difficult and the results were considered less reliable than those of the other areas.

In retrospect it is evident that in these two first years of exploratory surveys there was not sufficient time for identification and sampling to allow a quantitative analysis by species or groups of species of the composition of the main group: the small pelagic commercial fish. Such estimations must be based on elaborate sampling combined in a concurrent analysis with the echo integration data. However, it was worth undertaking further analyses of the catch data from these first exploratory surveys. Inherent difficulties in interpreting catches and catch rates in pelagic trawl hauls in terms of species composition of the pelagic assemblages must be kept in mind as these also reflect differences in behaviour and catchability of the various species.

In the following summary the catch data of all surveys made in 1975–76 were grouped by main production regions (Kesteven et al., 1981). Only hauls with the pelagic trawl containing catches of small pelagic commercial fish were included. Table 3.12 shows the data for the most abundant species from the east coast of Somalia. The assemblage was dominated by Clupeidae. The high incidence of round herring (Etrumeus teres) was assumed to be due to the relatively high catchability of this species. Indian oil sardinella (Sardinella longiceps) was one of the main species. Less abundant were several other sardinellas, and among the Carangidae, scads (Decapterus russelli) and horse mackerel (Trachurus indicus).

Table 3.12 East Somalia: Species composition (main species only) of catches in hauls with the pelagic trawl containing small pelagic fish

Large catches of mackerel (Scomber japonicus) were taken with the bottom trawl. This species with its partly bentho-pelagic behaviour seemed to be characteristic for the Somalia upwelling area and this may be related to the higher oxygen levels in deeper waters of this region as compared to the situation further north in the Arabian Sea (Wyrtki, 1971).

Table 3.13 shows the species composition in the somewhat more comprehensive sampling of the upwelling region off Yemen and the east coast of Oman. Some demersal hauls with a predominant catch of small pelagic fish were included. One haul of 12 t of Sardinella albella was excluded. This was an assemblage dominated by Clupeidae with Sardinella longiceps as the main species, especially off Yemen, but with a strong component also of Carangidae, with Trachurus indicus as a main species especially off Oman. The relatively low incidence but high catch of this species indicated a restricted distribution, but high local abundance.

Table 3.13 Yemen-Oman: Species composition of catches in hauls containing small pelagic fish, including all pelagic trawl and selected demersal trawl hauls with dominant catch of small pelagic fish

For Pakistan, additional information on the composition of the neritic pelagic fish became available in the special programme which followed the regional programme in January-June 1977 and included five surveys (IMR, 1978a). This more extensive sampling confirmed the earlier observations and showed an assemblage on this shelf different from that of the western upwelling areas (Table 3.14). The fauna was more varied with a larger number of species. The high incidence and catch of anchovies consisting of both Stolephorus spp. and Thryssa spp. was probably partly related to their high catchability, but they seem to have dominated, especially on the eastern Sind coast and Somniani Bay where the highest densities of small pelagics were found. The higher proportion of anchovies of the total small pelagics on the eastern Pakistan shelf as compared to the western upwellings is probably related to the wider inner shelf habitat of this coast compared to the narrow western shelves and to the effect of the discharges from the Indus river. Rainbow sardine (Dussumieria acuta), oil sardine (Sardinella longiceps) and Sind sardinella (Sardinella sindensis) were also common especially along the western Makran coast. The two scads (Megalaspis cordyla and Decapterus russelli) dominated the Carangidae. The Scombridae were represented by various Scomberomorus species which appeared with high incidence (about 30%), but low catch rates in the demersal hauls.

Table 3.14 Pakistan: Species composition of catches in hauls containing small pelagic fish, including all pelagic and selected demersal trawl hauls with dominant catch of small pelagic fish

Revision of biomass estimates of small pelagics

As discussed in Section 2.2, the likely systematic errors of acoustic survey methodology are expected to result in underestimates of biomass, particularly for fish occurring in dense schools from which instrument saturation would be common. It was not possible to adjust for these biases, but the high target strength level used during the first two years of survey -29 dB/kg (for 17 cm fish) had to be adjusted in order to obtain data comparable to those of later surveys. In the discussion of methods in Chapter 2 it was suggested to assume a level of -32 dB/kg for small pelagic fish. A decrease of the target strength (TS) by 3 dB implies a doubling of the biomass estimates from the first two years of survey. For the western coastal upwelling areas dominated by Clupeidae a mean fish size of roughly 17 cm was used. For the Pakistan coast the value of C, the factor converting integrator deflection to biomass had to be adjusted to the smaller size (about 8 cm) of the dominating anchovies (Table 14 in Kesteven et al., 1981). An approximate adjustment is achieved through halving the C factor.

The survey-to-survey variation of the biomass estimates for each sector shown in Table 3.11 was considerable. Part of this may have been caused by displacement of stocks or points of high production between sectors. The bulk of the fish in Yemen waters was thus most often found eastwards in a distribution which seemed to continue along the Oman coast. South Somalia, Socotra and the eastern part of north Somalia would also seem to form one production area. These areas would relate to the coastal upwellings off the Arabian Peninsula and north Somalia respectively. A rearrangement of the biomass estimates for these general production areas is shown in Table 3.15 where the data have been adjusted to a 3 dB lower TS level (from -29 to -32 dB/kg) for all areas and in addition with the Pakistan estimates adjusted for small fish size (halving the C factor).

There was a considerable variation in the series of data for each area, and there was no apparent seasonal cycle for the two upwelling areas (Somalia and Arabian Peninsula). For the area south and southeast of the Arabian Peninsula the last three surveys gave fairly consistent estimates of close to 1.5 million t of which the Oman coast accounted for about two-thirds or 1 million t. The high estimates for Pakistan from January to May may have been an effect of a seasonal production cycle in the short-lived anchovies.

Table 3.15 Biomass estimates of small pelagic fish by major production areas and surveys, (1,000 t) adjusted for lower levels of target strength (see text and Table 3.11)

Catchability

The availability of a resource for commercial fishing depends on its appearance in catchable concentrations. An evaluation of the catchability can be made by examination of the distribution charts to ascertain the parts of the total biomass recorded at various degrees of densities: such data from Surveys 3 to 6 for the different areas appear in Table 3.16. As discussed in Chapter 2 the general experience from the DR. FRIDTJOF NANSEN surveys was that densities higher than about 300 t/nmi² represent areas where small pelagic fish are aggregated in “fish areas” or “school areas” where many schools were found especially during the day and where at night the fish were often found in more or less continuous more dispersed layers. Fish in such areas are vulnerable to commercial fishing. Table 3.16 shows that most of the biomass of the small pelagics in the northwest Arabian Sea was found in densities which should make them available to commercial fishing by purse-seining.

Table 3.16 Proportion of biomass found at levels of higher densities (%)

The distributional characteristics of the small pelagics on the Pakistan shelf were entirely different. By far the main part of the biomass was found in scattered layers and small dispersed schools. School areas with higher densities were almost absent in the west on the Makran coast and Somniani Bay and were only found in the east on the Sind coast in the two spring surveys (1 and 3). Densities higher than 250 t/nmi² then only represented 9% and 17% respectively of the total biomass estimate. It is concluded that therefore these resources would not be a suitable target for purse-seine fishing.

Semi-demersal fish

In addition to the small pelagic schooling fish which represented the main objective of the surveys the reported data included acoustic records of targets recorded originally as demersal, fish from their appearance in layers, usually of low density close to the bottom or in the lower part of the water column, but for which semi-demersal would have been a better classification. Since many semi-demersal fish tend to rise off the bottom at night a distinction should be made between day and night recordings. This was done only in later surveys, and the averaging of day and night recordings in the first surveys reported here results in an underestimate of this category.

Special problems complicated the description of the fish species more closely associated with the bottom in this region. The monsoon cycle causes a seasonal intrusion of water with a low oxygen content onto the shelf during the southwest and post-southwest monsoon causing drastic changes in the distribution of demersal and semi-demersal fish (Bianchi, 1992a and b). As a consequence, a high variation in biomass estimates between different seasons can be expected. Another effect is that some fish normally considered as demersal, such as threadfin breams (Nemipteridae) and catfishes (Ariidae) have apparently adopted a more pelagic behaviour in this region.

Results of bottom trawling which was in general used for identification and sampling of the demersal fish component of the integrator contributions, was probably positively biased towards the true demersal species which were often reported to be the main component of the catches: groupers (Serranidae), croakers (Sciaenidae) and scavengers (Lethrinidae) and to some extent snappers (Lutjanidae), while it is likely that the acoustic mid-water targets may have been predominantly: threadfin breams (Nemipteridae), catfishes (Ariidae), hairtails (Trichiuridae) and ponyfishes (Leiognathidae). Ponyfishes were often found in considerable abundance and their echograms were probably correctly identified.

The very narrow shelf of East Somalia with its strong currents represented particular problems for fishing and hence for identification. It is thought that most of the recordings of semi-demersal fish in this region were not adequately identified and it must be assumed that they have consisted mainly of non-commercial species: porcupine fish (Diodontidae), lizard fish (Synodontidae) and small-sized cardinal fish (Apogonidae). Deep-swimming schools and layers of mackerel (Scomber japonicus) may also have been included. Bianchi (1992b) presented a study of the demersal assemblages on the continental shelves in this region based on bottom trawl surveys of later assignments.

Identification of the semi-demersal fish can thus not be based directly on an analysis of catch compositions since bottom trawl catches must be assumed to be biased. Furthermore, there was hardly any aimed fishing with mid-water trawls for these targets. However, a general identification based on knowledge of fish behaviour and on data from IMR (1977b), IMR (1978a) and Bianchi (1992a and b) is attempted in Table 3.17. There appears to be some difference in order of abundance, with ponyfish and threadfin breams being most common in Yemen and Oman, while hairtails and catfishes dominate in Pakistan.

A review of the biomass estimates of the semi-demersal group by surveys as presented in Table 12 in Kesteven et al. (1981) shows a considerable variation, which may partly be related to the seasonal changes in the environment. As most of the semi-demersals off the Somalia coast seem to have been non-commercial species, no estimate will be attempted for this region. For Yemen-Oman the hydrographic profiles show insurgence of low-oxygen water up to about 50 m depth during the post-monsoon season (surveys 3 and 6) which leaves estimates of 266,000, 46,000 and 137,000 t respectively for the three pre-monsoon surveys 1, 4 and 5 (Kesteven et al., 1981). For Pakistan only Survey 5 represents a pre-monsoon coverage where the distribution of integrator readings suggests a fairly consistent identification of the demersal group. An estimate by area integration gives about 220,000 t. Table 3.18 shows the mean biomass estimates from the above selected surveys referred to a target strength level of -32 dB/kg (for 17 cm fish). It should be remembered that the assessment of semi-demersal fish was not a specified objective of the survey programme, and considerable reservations should be attached to these estimates. More confidence could perhaps be afforded to the combined pelagic/demersal estimates as representing a measure of the “total acoustic biomass”.

Table 3.17 Main contributors to the recordings of the semi-demersal group by production areas in approximate order of appearance in the catches

Table 3.18 Biomass estimates of semi-demersal fish based on acoustic observations in selected pre-monsoon coverages (t)

The special surveys of the more important parts of the northwest Arabian Sea undertaken later in the programme included bottom trawl surveys (swept-area estimates) which provided more comprehensive and reliable information on semi-demersal and demersal fish (see Section 3.5).

Mean biomass densities

The mean standing fish biomass per unit area may be a rough indicator of productivity. In the northwest Arabian Sea the distribution of the small pelagic fish extended in many cases beyond the narrow continental shelf and the survey area was in an approximate way adjusted accordingly, being defined as inshore surveyed space. The mean biomass density was referred to this area, but for general comparisons a reference to the continental shelf area was also made (Table 3.19). The estimated mean densities for the small pelagics were similar for the two upwelling areas, Somalia and the Arabian Peninsula, but about 50% lower for the Pakistan shelf.

Table 3.19 Mean biomass densities (t/nmi²) in main production areas referred to inshore surveyed space and continental shelf

3.4 SURVEYS OF MESOPELAGIC FISH IN THE NORTHWEST ARABIAN SEA, 1975–84

Survey effort

One of the main findings of the DR. FRIDTJOF NANSEN surveys in the northwest Arabian Sea was the wide distribution and high abundance of mesopelagic fish. This result could perhaps have been foreseen.

The faunistic regimes of coastal upwelling systems are known to contain a shoreward assemblage of Clupeidae, Carangidae and Gadidae and an oceanward assemblage of mainly Myctophidae and Scombridae (Cushing, 1971b). The highly productive zones in the northwest Arabian Sea were observed to extend much further offshore than in the coastal upwelling zones of eastern boundary currents. Recordings of mesopelagic fish had already been made in the area, viz., those reported by the RSS DISCOVERY's participation in the IIOE (Anon, 1963). They were not at the time identified as mesopelagic fish, but this could have been deduced from the description of their distribution and behaviour. This was the only fish observation reported from the whole of the five year IIOE programme. There is now also evidence from otoliths in deep-sea cores of long-term prominence of Myctophidae in the region (Anon., 1991a).

Mesopelagic fish was an important element in the findings from the following assignments of the DR. FRIDTJOF NANSEN programme:

Main exploratory fish surveys:

February 1975-November 1976, 5 coverages northwest Arabian Sea
January-June 1977, 5 coverages Pakistan

Special surveys for mesopelagics:

July-August 1979, Gulf of Oman and Gulf of Aden
January-February 1981, Gulf of Oman and Gulf of Aden
February-March 1983, Gulf of Oman
September 1983 and May-June 1984, Iranian part of Gulf of Oman

The results regarding mesopelagics from the main fish surveys are described by Gjøsæter (1977, 1981b) and reviewed in Gjøsæter and Kawaguchi (1980). The Pakistan 1977 surveys are reported by IMR, 1978a. Of the special surveys for mesopelagics the 1979 survey is reported by Gjøsæter and Myrseth (1979) and also together with the 1981 survey in Aglen et al. (1982). The February-March 1983 survey is reported by Gjø;sæter and Tilseth (1983) and the fishing trials by Schärfe (1983). The 1983–84 surveys from Iran are reported by IMR (1983c) and Ona (1984a).

A summary follows of the findings of the main surveys 1975–77 regarding the distribution, composition and behaviour of the mesopelagics in this area and the reported biomass estimates. Since the biomass estimates of mesopelagics were very high and the catch rates obtained in some areas were promising there was a potential commercial interest in these resources. Special surveys were therefore mounted in the high density areas to provide more detailed biological and technological information; their results are reviewed below. They are of special interest as they were planned to resolve specific problems relating to these fish and were conducted by scientists who had by that time acquired considerable knowledge of their biology and behaviour.

Results of the main surveys, 1975–77

The intensive coverage of the shelf areas usually included the slope which forms the inner limit of the distribution of the mesopelagics, but the density of the course tracks in the central parts of the Gulfs of Oman and Aden and the open ocean was less, as can be appreciated from Figure 3.12 (Section 3.3) which shows the course tracks of one of the general surveys. The oceanward extent of the survey had been adjusted to correspond in a very general way with the seaward limit of the highly productive part of the western Arabian Sea. No attempt was made to adjust the survey area to the distribution of the mesopelagic fish.

The area from Somalia to the Pakistan-India border was covered five times during 1975–76 and the Pakistan shelf was covered five times during January-June 1977.

The acoustic instrumentation (38 kHz echo integration system) allowed observations within a 500 m depth range. Fishing for identification and sampling was made with a 1,360 mesh mid-water trawl.

During the day a deep-scattering layer named D2 layer was found throughout the area between about 250 and 350 m. In the northwestern part and occasionally in the Gulf of Aden, an additional layer D1 was found at 100–200 m. The depth of the layers changed at night when all or part of the D2 layer would migrate upwards and join the D1 layer, if existing, and form a near-surface layer, N1, at 10–100 m depth. Sometimes part of the D2 layer would remain in deep-water also during the night forming a layer named N2. Figure 3.14 which is based on acoustic observations from a fixed station in the Gulf of Oman in 1976 illustrates this diurnal behaviour. The vertical migrations were fast, the downward migration at sunrise often being completed within 20 min (Ona, 1984a).

Figure 3.14 Vertical migrations of mesopelagic fish as observed during a diurnal station. (1) Schools and very dense aggregations, (2) dense recordings and (3) scattered recordings (from Gjø;sæter, 1981b)

Many species were identified from the whole area, but Benthosema pterotum dominated most catches in the high density areas in the northwest and this was the only species found in the area of the fixed station in the Gulf of Oman. Benthosema fibulatum and Diaphus spp. were also common. The size compositions (total length) in samples of B. pterotum ranged from 20 to 50 mm with most means between 30 and 40 mm. The Benthosema species were assessed to have a life cycle of one year or less.

Biomass density was estimated from the acoustic integration data. High densities were often encountered in the slope or adjacent areas, but analysis of observations further offshore in selected profiles of 150–200 nmi from different parts of the total coverage showed no general trends of decrease or increase in a seaward direction. There was thus, as shown in Figure 3.15, a very wide total distribution of the mesopelagics in contrast to other fish populations usually found in smaller contagious distributions.

The sub-divisions used in the analysis are shown in Figure 3.16. The mean biomass densities varied between areas in an at least partly consistent way as shown in Figure 3.17, with area A, the Gulf of Oman showing the highest mean densities in nearly all the surveys. Relatively high densities were also found in the Gulf of Aden in the first two surveys while in the last two surveys the density here was low, and the species composition had changed.

Figure 3.15 Mean integrator deflection of mesopelagic fish (Autumn 1975 survey)

Variations in catch rates between surveys seem to confirm that the observed variation in mean density was at least in part caused by changes in population density. Variance caused by the applied survey methodology may also have affected the data. Thus the persistent high observations during the first two surveys in nearly all areas may indicate some bias in the first survey, e.g. inadequate acoustic identification and sampling. The observations from this first survey will thus be omitted in further analysis of densities and abundance although they are included in the first historical account of the findings. In some surveys coverage may have been inadequate in certain areas.

Mesopelagic fish was also included in the five fish assessment surveys off Pakistan in January-June 1977. The biomass estimates of the first three surveys were somewhat higher than in 1975–76. Declining levels in the last two surveys off Pakistan were probably an artefact caused by a reduced performance of the acoustic system. Catch rates were generally low; only 9 of 33 hauls exceeded 400 kg/h and only 4 exceeded 1 t/h.

The estimated densities of mesopelagics observed in the 1975–76 surveys were used to calculate total biomass by surveys which was found to range between 56 and 148 million t for the whole area covered, with a mean of 100 million t. These were noted as remarkably high levels of fish biomass and their veracity has repeatedly been discussed and questioned since the first summary report by Gjø;sæter (1977).

Figure 3.16 The investigated area and its subdivisions		Figure 3.17 Mean integrator deflection for sub-divisions shown in Figure 3.16 and by surveys (from Gjø;sæter, 1981b)

The results of fishing experiments also demonstrated the existence of a considerable quantity of mesopelagic fish. Comparisons with estimates of density calculated from catch rates in fishing trials on selected good recordings of mesopelagics showed figures exceeding the mean acoustic density estimates. Some 26 stations gave catch rates higher than 400 kg/h of which 18 exceeded 1 t/h, of which six with 5 t/h and a highest rate of 20 t/h. The locations of these fishing stations cover, as shown in Figure 3.18, a very large area but indicate a restriction of high densities of fish to the Gulf of Oman and to the shelf break areas outside the Gulf of Oman. Catch rates in the open ocean were very much lower with no indication of possible commercial rates. The very high total biomass estimates were therefore primarily of interest in a context of total production and trophodynamics although this was perhaps not recognized at first.

Based on primary production estimates presented by Cushing (1971a) for the area covered by the survey, Gjø;sæter (1977) found that an annual assumed production of 100 million t of mesopelagics represents between 1 and 2 % of the primary production and about 10% of the secondary production. If an ecological efficiency of 10% is assumed at each trophic level, the mesopelagic fish would utilize the entire secondary production.

There are, as discussed below, reasons for suspecting the existence of a bias in these biomass estimates which may have caused a considerable overestimate, perhaps by a factor of 2–3. This, however, does not change the general finding of the surveys that the mesopelagics were by far the most abundant type of fish in the northwest Arabian Sea and that they represented a major component of the ecosystem.

Figure 3.18 Trawl stations in the 1975–1976 surveys giving more than 400 kg/h of mesopelagic fish, catch rates in t/h (adapted from Gjø;sæter, 1981b)

Results of the special surveys, 1979–84

Further survey work on mesopelagic fish was restricted to the areas of highest observed densities: the Gulfs of Oman and Aden. The objectives were to: obtain new acoustic abundance estimates; provide other evidence of densities; collect data for further biological and ecological studies, and carry out experimental fishing.

In the special survey mounted in July/August 1979 the acoustic abundance recorded in the Gulf of Oman was found to be about 8 million t, which was within the range observed previously, but in the Gulf of Aden the biomass was estimated at only 4 million t. This low level compared with previous findings was confirmed by low catch rates.

An attempt to use mean catch rates from the main scattering layers for the estimation of total biomass gave proportions between 20 and 50% of the mean acoustic estimate assuming that the trawl would catch all fish passing through the trawl opening.

In January/February 1981 most of the work was conducted in the Gulf of Oman. Studies of the various scattering layers were made in repeated surveys of the main part of the Gulf and in six detailed surveys of a selected area.

Echo integration of a depth interval corresponding to that covered by the trawl opening during towing provided a set of observations for estimating apparent trawl efficiency: the ratio between catch/m³ filtered water and acoustically-measured density. This was found to be very low for the D2 layer. A possible explanation given for this was the lack of a herding effect of the bridles and trawl wings at these depths.

Experiments testing the effect of underwater light on the mesopelagics were also conducted and underwater photographs of various layers were taken.

The 1981 biomass estimate for the total Gulf of Oman was 11 million t (also within the range of previous observations) and high catch rates were obtained. The Gulf of Aden estimate was higher than in 1979 at 16 million t, but the catch rates were very low and a change in species composition seemed to have occurred.

The February/March 1983 survey in the Gulf of Oman had various objectives: to continue the studies on behaviour and biology; to monitor the biomass of the mesopelagics, and to undertake a series of experimental fishing trials with gears specially designed for this purpose.

Two coverages gave biomass estimates of 8.0 and 4.7 million t, a rather high variation. An analysis of the contribution from the different types of scattering layers now showed significantly lower night than day levels. An examination showed that this was the case also for the 1981 data similarly analysed. The D2 layer gave especially high echo intensities. An observation of importance for the commercial fishing experiments to be carried out this season was that the mean fish density in the D1 layer was considerably lower in 1983 than observed in 1981.

The fishing experiments reported by Schärfe (1983) were conducted over an 11-day period with mid-water trawls adapted with enlarged forward extensions of large mesh netting, 800 mm and 1,600 mm stretched, which increased the opening of the trawl from about 250 m² to respectively two and three times that area.

The results are summarized in Table 3.20. Herding by the large mesh net must have been effective in the D1 layer as demonstrated by good overall catch rates there. Fishing was also effective with artificial light in the N1 layer. The results were described as successful and promising especially in view of the observed low general density of the D1 layer encountered in the trials as compared with the density found in 1981.

The densities estimated from acoustic integration during 27 hauls on “reasonably good” D1 layers in 1983 ranged from 20 to 412 g/m² with a mean of 80 g/m² corresponding to 0.7 to 24.2 g/m³ with a mean of 4.3 g/m³. It should be noted that these represent high densities in locations selected for fishing. On the other hand, the mean estimated density of fish in the D1 layer in the 1983 main surveys was 0.4–0.5 g/m³ corresponding to 10–12 g/m². In the 1981 survey the D1 layer density was estimated by Gjø;sæter and Tilseth (1983) to be about 5 times higher with 2.4 g/m³ and 59 g/m².

Table 3.20 Summary of fishing trials on lanternfish (adapted from Schärfe, 1983)

The D2 and N2 layers were, however, reported by Schärfe (1983) to be “completely useless for commercial fishing”. The comments of Schärfe as an experienced fishing technologist after these tests are worth noting. The acoustic estimates of fish density made by integrating the layer in front of the trawl allowed testing of the efficiency of the gear. A very low apparent efficiency was experienced in the D2 layer causing Schärfe to suggest that further similar experience might make a reconsideration of the “conversion factor” for the D2 layer desirable.

The two last special surveys on mesopelagics covered the Iranian part of the Gulf of Oman in September 1983, IMR, 1983b and May-June 1984. The recordings were reported by Ona (1984a) to be generally scattered to dense, but belts of high density were found. The D2 layer, below 200 m was reported to give blurred recordings, but very high integrator values.

The reports, IMR (1983c) and Ona (1984a) show the distribution of mesopelagics by relative densities. The results of biomass estimates based on an assumed target strength of -34.2 dB/kg (17 cm fish) were as follows:

September 1983: 2,384,000 t in an area of 10,589 nmi²
May-June 1984: 2,600,000 t in an area of 13,743 nmi²

The distributions cover the deep-sea areas of the Gulf of Oman from the edge of the shelf to the mid line, from the border with Pakistan at about 61°30'E to Hormuz Strait. Densities seemed to increase somewhat westwards.

These are the only estimates from special mesopelagic investigations of the Iranian part of the Gulf of Oman. In the September 1983 survey the acoustic system was still the EKS echosounder with the analog QM integrator. In the May-June 1984 survey the new EK400 system had been installed with the digital QD integrator. Both systems had been well calibrated. It is of interest to compare the densities observed with those obtained in the Oman part of the Gulf. The closest survey in time is from February 1983 and covers the Gulf eastward to about 59°20'E. Table 3.21 shows a comparison of densities in the various parts of the Gulf from these surveys.

Table 3.21 Observed mean densities from parts of the Gulf of Oman in the 1983 and 1984 surveys (g/m²)

Compared with the mean densities in Oman waters in February 1983 the density of mesopelagics in the whole of the Iranian EEZ up to the Hormuz Strait was about 50% lower in September 1983 and May-June 1984. In the Iranian part of the Gulf proper the densities were about 25% lower. It is uncertain whether these comparisons made at about six months intervals are valid. They suggest that extrapolated estimates of the biomass of the whole Gulf, made on the basis of observations from the Oman part only such as those from January-February 1981 and February 1983, may be about 12% too high.

The Iranian 1983 and 1984 surveys included only a few fishing trials for the purpose of identification and size sampling. Various programmes of commercial fishing trials have later been conducted in Iranian waters, but the results are not generally available. A 1992 programme with a powerful mid-water trawler yielded high commercial rates in a limited area in the western part of the Gulf according to Beltestad (1992).