Fish Stocks in the Arabian Sea

If we take the main interest of U.S. GLOBEC to be secondary production and its transfer to higher trophic levels, then the Arabian Sea offers a first-class problem. The northern Arabian Sea is the habitat of large mid-water fish stocks (Gjosaeter, 1984). These stocks reside all along the outer edges of the coastal zone of the Arabian Peninsula, off Pakistan, and the Gulf of Oman. Doubtless there are also sizeable stocks off Somalia and northern India. Off the Arabian Peninsula and in the Gulf of Oman these stocks are dominated by myctophids, mostly by Benthosema pterotum, although Benthosema fibulatum, Diaphus arabicus (see Kinzer et al., 1993), Myctophum spinosum, and Symbolophorus evermanni are occasional large contributors. Possibly the B. pterotum population is the largest, localized fish stock in the world, amounting to 100 million tons! It has been suggested that this very large stock of the one species derives from the very small stocks of all other fish; for some reason B. pterotum and other myctophids are the competitive dominants. There are references in the literature to the importance of the oxygen minimum layers in this, but there is no suggestion of a specific mechanism.

A good deal is known about the biology of B. pterotum. Gjosaeter and Tilseth (1983) reported a stomach content study, which showed that copepods, and mesozooplankton generally (euphausiids, Oikopleura, etc.) are the diet. 94% of stomachs with food contained copepods. In a separate study, Delpadado and Gjosaeter (1988) showed that copepods make up 35 to 55% of the diet by weight, with the remainder varying among sampling sites and times (Fig. 15). The numerically dominant copepod prey were small species, Corycaeus sp. and Oncaea sp. Since these are believed to be associates of gelatinous zooplankton, it is possible that these myctophids take their meals at these "feeding stations." However, most of the stomach content mass was "larger calanoid species (prosome lengths 0.8 to 2.5 mm), of which the dominant species were: Acartia danae, Candacia sp., Centropages furcatus, Clausocalanus sp., Eucalanus tenuis, Euchaeta marina, ...." In an earlier study, Gjosaeter (1981) showed that feeding mostly occurs early in the night, slowing in the late night, and is minimal at depth in daytime. In a study in the Red Sea, Delpadado and Gjosaeter (1987) showed that mean copepod prey size increased with increasing length of B. pterotum. Kubota (1982) sampled this species in Suruga Bay and found that its principal diet is copepods (Calanus, Oncaea, many others) and euphausiids (Euphausia sp.), the latter especially important seasonally. Thus, this fish must be in the middle of a fairly short food chain: diatoms-copepods-B. pterotum-larger fish, squid, marine mammals. Outside of the upwelling season, and perhaps generally in the open sea, the food chain may well have an additional step: flagellate phytoplankton-protistan grazers- copepods, since the region has an unusually high coccolithophorid contribution to the production (Codispoti, 1991). We are now well prepared to study pelagic food webs of either type.

Benthosema pterotum is associated with the continental slope all around the northern Indian Ocean and in the western Pacific as far north as Japan. There are populations in the Gulf of Panama of the very close relative Benthosema panamense. Thus, it has a tropical-subtropical, Indo-Pacific distribution, but it is not found over oceanic depths. The exact effects of topography in determining its distribution are not fully understood. Probably there are strong ecological analogies with the large midwater fish biomass of the Benguela upwelling system (Prosch 1991; Armstrong and Prosch 1991). It is readily collected by large trawls (or even small trawls) so that diet, growth rates, biochemical composition, and many other aspects of its biology can be studied readily. It is an excellent sonar target, so stock size can be estimated by sonar survey. It is a strong diel migrator, inhabiting surface layers at night, and only descending to 150-350 m during the day. Gjosaeter (1981, 1984) showed with sonar that the population in daytime is divided into two layers, one at 150-200 m, the other at 250-325 m, with a distinct gap between (Fig. 16). Samples from these separate groups were not distinct in size, reproductive state, sex ratio, or anything else that Gjosaeter could examine. Gjosaeter suggests that its distribution is related to and controlled by the shallow oxygen minimum layer of the Arabian Sea; the fish apparently are driven inshore by intensification or elevation of the minimum. The vertical gap in daytime distribution in the Gulf of Oman coincides with a local O2 minimum (ca. 1 ml L-1) there.

Benthosema pterotum has an annual life history with reproduction at age approximately 7 months (Gjosaeter, 1984). Egg production is observed in females ranging from 27 mm (300 eggs) to 52 mm (3,000 eggs) (Delpadado 1988, Hussain and Ali Khan 1987) with very little variation among sites (Fig. 17). Despite the wide range of size at reproduction, Delpadado (1988) thinks that there may be only one round of egg production with death ensuing, since apparently-spent females are found across the same size range. The presence of three size classes of eggs in ovaries of all sizes of fish (Hussain and Ali Khan 1987), however, makes a single reproductive burst seem unlikely. At any rate, the life span is surely no more than a year, which implies that the entire 100 million tons is produced every year! This is a staggering sum, implying a primary productivity around the arc of continental slope from Cape Guardafi to India on the order of 1 billion tons carbon. Measured primary productivity in this area is indeed high, often running 2 g C m-2 d-1, and sometimes reaching 6 g C m-2 d-1 (Codispoti, 1991), but these high figures do not prevail during the entire year. GLOBEC work in the region can benefit from association with U.S. JGOFS, drawing insight from proposed U.S. JGOFS primary production studies.

Gjosaeter and Tilseth (1988) have shown that B. pterotum in the Gulf of Oman spawns at depths of 100 to 300 m early in the night. The slightly buoyant eggs hatch within 12 hours before they reach the surface. The larvae are well described (Tsokur 1982). Some further work on vertical distribution and diet of larval phases would be informative. Nellen (1973) states that larvae of B. pterotum overwhelmingly dominate collections along the coast of Pakistan, and later data (Ali Khan 1976) confirm that for areas off the edge of the shelf. Over the shelf off Karachi Sardinella sinensis is more abundant. There have been no larval surveys along the Arabian coast, but since adult B. pterotum are abundant, and larvae are abundant both in the Gulf of Aden and off Pakistan, it certainly is the dominant among larval fish all around the north rim of the Arabian Sea. No seasonality in larval stocks has been reported, although there are hints in the literature of greater larval abundance in the NE monsoon. Sanders and Bouhiel (1982, seen in abstract) considered recruitment to the fishable stock to be confined to the two monsoon transition seasons. The basis for this statement is not given in the available abstract. Venema (1984) shows some seasonality in total pelagic and demersal fish stocks in various coastal sectors of the northern Arabian Sea, and this must certainly derive mostly from variations in pelagic species. Stocks are high in the NE monsoon, lower in the SW monsoon. The data are sketchy, with a glaring hole in the middle of the SW monsoon. Delpadado (1988) found mature and spawning fish in all seasons, and found no definite evidence of any seasonality in population processes. It is, however, very likely that the life history of B. pterotum is in some fashion in phase with the monsoon cycle. This aspect of coupling between the animal and its habitat could be a research topic.

Benthosema pterotum is a good subject for a direct study of larval recruitment and year-to-year variation in stock size and success, although spawning is not seasonally focussed enough for characterization of years by simple one-shot surveys. U.S. GLOBEC workers should be able to couple observations from primary productivity through tertiary productivity about as well as for any region--taking this dominant fish as a key study subject, an upwelling system component suitable for detailed population quantification. The main drawback to this species as a focus is that it tastes terrible. According to Kubota (1982), fishermen in Suruga Bay who eat large quantities of Diaphus spp. sort out and discard B. pterotum as inedible. That does not mean that this huge production is useless; fish oil and protein have other uses than direct human consumption. Studies in India (Gopakumar et al. 1983; Nair et al. 1983) show that meal and hydrolysate from B. pterotum are excellent protein supplements in fish and poultry feeds. These myctophids are readily fished; Norwegian results reached 100 tons hr-1 with a sonar-guided, 750 m2 (15 X 50 m) double warp trawl (which is a seriously large piece of gear).