5. TRAWLING TRIALS AND RESULTS

5.1 Fishing Conditions
5.2 800 mm Meshsize Trawl
5.3 1600 mm Meshsize Trawl
5.4 Trawling with Light

Since the purpose of the trials was to assess the possible economic feasibility of mid-water trawling for B. pterotum, most of the fishing was done in a simulated commercial way, with only a few tows for initial technical tests and adjustments and a few sampling tows in other than D1 layers. Consequently, for most of the cruise, the main emphasis was concentrated on harvesting the D1 layer, which offered by far the best concentrations. After a few trials on the D2 layer (tows 1, 12) and the N2 layer had given most disappointing catches and the yield of sampling with the Harstad trawl on those and also the N1 layer had confirmed that they were not concentrated enough for commercial size catches, it was the impression that the D1 layer was the only worthwhile target. That would have meant that only about 12 hours per day were available for commercial fishing. Only after the catch of tow, 47 had shown that night fishing with light on the N1 layer can yield commercial size catches, did the feasibility of round-the-clock trawling became apparent. It does not need to be stressed that this fact significantly increases the favourable aspects of a possible future commercial fishery. A 24 h test fishing towing back and forth without any searching for fish but just taking what happened to be available (tows 47 to 54) yielded in 15.1 fishing hours a total catch of 40.0 t, of which 15.8 t were caught with light from the N1 layer. Although this is not enough for actual commercial fishing, it is a handsome result for a first try.

A record of the trawling trials is given in Appendix 4 and a summary is tabulated in Section 1. As mentioned in Section 4, when looking at these tables, it must be kept in mind that commercial type midwater trawling usually requires frequent changes of towing power to aim the trawl at the densest concentrations. This is also true for fishing layers of the D1 type and even, to a lower degree, the more uniform N1 layer. The unavoidable changes in towing speed, as well as shape and size of the net opening, prevent the determination of accurate average values of these parameters and also affect the accuracy of the values (e.g., distance towed, volume filtered, etc.) derived from them. The most reliable data are total catch and catch per hour, which are of predominant interest for commercial considerations. The judgement of the accuracy of acoustic stock assessment and the values derived therefrom is up to the individual reader.

For a first reasonably serious attempt at commercial type midwater trawling, a total catch of about 230 t in 61.4 fishing hours, with an overall average of 3.8 t/h and an overall average fishing efficiency of 38 percent of the fish in the path of the net opening caught is certainly quite encouraging.

First of all this result proves the technical approach of guiding and shepherding B. pterotum with large meshes at low towing speed to be correct. Since the catching efficiencies of both versions, the 800 mm and the 1600 mm maximum meshsize, were found to be virtually the same for reasonably good D1 layers, a further increase in meshsize may enable less towing power or still larger trawls. Also the gradual decrease in meshsize in the belly may very well offer scope for additional reduction in drag. Further trials with improved net designs are clearly indicated. Also lower and higher towing speeds should be tested systematically to better determine the most economical range and in particular its lower limit.

An analysis of the catch rates confirms the general impression that the D1 layer is the primary target, with best catches shortly after daybreak and before sunset. With better fish abundance, as was found during previous surveys, day fishing on D1 layers could very well be sufficient to catch the 100 to 150 t which are considered necessary for a medium size trawler to make some profit (see particularly tow 58).

The discovery that artificial light can be used to significantly increase the catch rates of night trawling in the N1 layer considerably increases the chances of commercial operations, even if the catch rates should remain lower than for daytime trawling. It is expected that more efficient lighting arrangements could largely improve night trawling and systematic trials are clearly indicated.

5.1 Fishing Conditions

According to a preliminary report of the Cruise Leader, Prof. J. Gjøsaeter, the total biomass of mesopelagic fish was slightly less in 1983 (about 9 million t) than in 1981 (about 11 million t) and also lower than the average found for 1975-76. But what was even more disadvantageous for the trials was the unexpected situation that (quoted):

“The most striking difference between 1983 and the previous years is, however, the distribution of D1. In spring 1976 and in spring 1981, for which detailed data are available, the D1 layer was rather constant all through the day and over most of the area. In 1983 a D1 layer usually formed during the morning migration but, after a short period (often one hour or so), it often descended and mixed with the D2 layer. Often it did not reappear before the afternoon (1500-1600 h), and it increased in density till it migrated to the surface in the evening. This made the period during which D1 could be fished short compared to the previous years. And even during the period when D1 was present, it was never as dense as the densest concentrations observed in 1976 and 1981.”

The distribution and migrations of B. pterotum in the Gulf of Oman are comprehensively covered by previous cruise reports and relevant papers and the particulars of the trial period will be reported by the cruise leader. These reports include representative echo-grammes and a duplication in this report would, therefore, be redundant.

From a fishing point of view, earlier experience was confirmed, according to which the D2 and N2 layers, which extend from about 250 m depth down, are too thin in fish density to yield commercial size catches. The fact that catching efficiencies of only 8 percent and less were obtained (tows 1, 12 and 15) could indicate too high integration values, which could be due to other small organisms such as squid and crustaceans which were found in the meshes. Tow 8, which aimed at fluffy traces giving very high integration values but very little catch from the upper border of the D2 layer, probably belongs to the same category. Further similar experience may make a reconsideration of the conversion factor for the D2 layer desirable. On the other hand, very thin D1 traces with low integration values gave catching efficiency values of more than 100 percent (tows 4, 5, 7 and also 9, 32 and 43) which would indicate that thin traces are undervalued by the present integration conversion. (Depending on one’s view of acoustic stock assessment, it could be argued whether the accuracy of this technique is adequate for determining the catching efficiency of fishing gear or whether midwater trawl catches could serve for determining conversion values, but such an argument is certainly not intended here.) The acoustic stock assessment for comparative judgement of fish density, likely catches and catching efficiency of different trawl types was of great value for these trials and it is considered most desirable that the vessel of the pilot project should be provided with respective equipment and staff.

The fish density in the D1 layer was far from uniform and the layer was also not continuous. For finding denser spots the SIMRAD searchlight sonar proved to be quite useful within a limited range of about 300 m. With a transducer tilt of about 20 to 25 degrees and an automatic search mode of about 60 degrees to both sides, aiming at concentrations outside the ship’s course is possible and was practised. With higher fish abundance, the range would be larger, particularly over deep water, and the project vessel should, therefore have a medium to large size searchlight sonar to profit from this searching possibility and to further study its value for commercial fishing.

During these trials, the D1 layers were mainly found above and just outside the edge of the continental shelf, i.e., within visual or radar range of the coastline, enabling terrestrial navigation. Since, in former surveys and in other areas, the mesopelagic stocks were also found further out in the open sea, exact and convenient spot plotting without landmarks is highly desirable for finding the way back to good concentrations. The project vessel should, therefore, have satellite navigation.

An up-and-down netsonde is indispensable for midwater trawling over deep water for aiming the trawl on the fish and for counteracting possible fish movements in, below and above the net opening. It is also considered advantageous to steer the net so that the main concentrations are in the centre of the net opening to reduce the chance for escape through the large meshes. During these trials, netsonde observations of the fish in and outside the net opening did not indicate a distinct flight reaction (distance) of the fish from the netting of the lower panel and the footrope. In view of the amount of catch obtained and the overall catching efficiency of about 45 percent on reasonably dense D1 layers, the consultant has no doubt that the large mesh netting in the front part of the net had a guiding effect. It appears that the fright distance the fish keep from the netting is so small that it is obscured by echoes from outside the centre of the sound beam. This could be studied by means of a very narrow beam netsonde during future trials.

Fish behaviour towards artificial light will be discussed below.

In summing up, it must be stated that the close cooperation with fishery biologists and acoustic stock assessment specialists was most useful and the same or similar arrangements would equally benefit future trials.

5.2 800 mm Meshsize Trawl

The fishing trials were started with the combination of the major part of the belly of the krill trawl with an enlarged front part of maximum 800 mm meshsize stretched (800 mm trawl, Appendix 7) because this version was assumed to provide the best chances for success. The reasons for this assumption were the not too excessive maximum meshsize, the reasonably well balanced decrease in meshsize in the belly, the small meshsize in the aft part of the belly and the comparatively long small meshed codend. Thanks to the excellent drawings explaining the joining of the pieces provided by the netmaker (H. Engel & Co., Bremerhaven), the fishing master of the DR FRIDTJOF NANSEN had this version ready when the consultant arrived in Dubai.

Since the reversion to the original krill trawl would have cost time and created frustration, the original plan to start the trials with the original trawl so as to have a standard for comparison with the performance of the trial trawls was given up.

The trawl was rigged as described under Section 3. The trials started with front weights of 520 kg each and, at a towing speed of 2.5 to 2.9 knots with 900 m of warps, the net opening measured 20 to 22 m in height and 30 to 36 m in width, giving an area of 660 to 720 m² in a fishing depth of 320 to 340 m. As can be seen by comparing with the table in Section 3 above, the opening height was according to expectations, while the opening width and consequently the opening area were considerably larger. This “overstretching” of the net was partly due to the comparatively high towing speed and the respectively strong sheer of the otter boards and partly to the large warp length with lower counterforce to the otter board sheer.

At the lower towing speed of about 2 knots and with reduced front weights (230 kg each), the shape of the net opening was reversed, i.e., opening height 33 m and opening width 25 m, giving an opening area of about 800 m² at a warp length of 300 m and a fishing depth of about 185 m. Since the D1 layers were mostly found in depths from 100 to 150 m, still shorter warps had to be used (mostly 250 m), which further impaired the opening width. A further reduction of the front weights was, nevertheless, considered not advisable with regard to shooting and hauling and the modest distortion was accepted.

As can be seen from Appendix 4, the average towing speed was somewhat higher (2.1 kn), favouring the opening width at the expense of the opening height. For tows 4 to 20 with reduced front weights, the average opening size was about 26 by 26 m = 675 m² which, in comparison with the table in Section 3, is quite acceptable. The comparatively higher values prove that the large meshes in the front part were well open and the size of the net provided by the design was fully utilized.

As can be seen from Appendix 4, there were considerable variations in opening shape and size which result from changes in towing speed and warp length required by any aimed midwater trawling and have to be accepted. The flexibility of midwater trawls can cope with these sometimes very drastic changes provided that the netting material has sufficient elasticity.

The crucial question of whether the large mesh netting in the front part guides the fish or not can be considered by comparing the amounts of fish in the volume filtered by the full net opening and by the smaller opening at the front edge of a smaller mesh net section further aft in the belly. For the calculation of the net cross section further aft, it is assumed that the net has the shape of a funnel from bosom to aft edge of the belly and that the cross section is a circle. For the actual length of the belly, the shape of the meshes is assumed to be in accordance with a hanging ratio E = 0.50, i.e., it is 20 percent less than stretched length.

Applying these assumptions to the 800 mm trawl with the average opening found when fishing for reasonably good D1 traces (Appendix 4) and choosing the seam between the large mesh front part and the remaining krill belly starting with 120 mm meshsize stretched as the position of the section further aft, the calculation gives the following results:

	Full net	Aft cross section
Average opening area (m2)	600	180
Volume filtered (m3 × 106)	25.0	7.5
Fish in volume filtered (t)	91.2	27.5
Actual catch (t)	41.2	41.2
Catching efficiency (%)	45	150

While a catching efficiency of 45 percent of the fish in the path of a trawl is generally considered to be realistic, a catching efficiency of 150 percent could only be explained by a very grave underestimation of fish abundance by the scientific acoustic assessment technique. This consultant prefers to believe in a reasonable guiding effect of the large mesh netting. The same consideration would apply to the 1600 mm trawl with its even longer large mesh front part (Fig. 6). After all, also 120 mm meshsize stretched would physically not be a serious obstacle for fish of only about 35 mm length.

The original small mesh codend with larger mesh cover and without ring strops proved to be too weak for handling larger catches. It burst with an estimated catch of about 4 t (tow 10) when being hauled up the chute. It was repaired by cutting off the damaged part of about 6 m length and by replacing the aft part of the cover with a bottom trawl codend available on board. Three ring strops were also provided to keep the codend together and thus reduce the pressure on its aft end. This improvised codend lasted for the rest of the trials and catches of up to 13 t (tow 58) were hauled on deck. Otherwise no serious net damage occurred.

As already mentioned in Section 3, the towing power of the vessel was more than ample for towing the 800 mm trawl with the trial speeds of up to 2.9 knots.

As regards the catches, the 800 mm trawl was somewhat at a disadvantage as compared with the 1600 mm trawl because of the smaller number of tows. Due to initial technical tests and trials in unrewarding layers, the number of tows useful for judging the catching efficiency was further reduced to the ten tows made on reasonably good D1 layers.

In view of the short time available and the uncertainty about the development of the trials with the other version, the testing of this net was terminated after the very successful tow 20 which, together with the proceeding tows, were considered to give a sufficiently clear idea about its technical and catching performance.

5.3 1600 mm Meshsize Trawl

The switchover from the 800 mm to the 1600 mm trawl (Fig. 6) was done overnight so that no day fishing time was lost. The composition of this trawl which, in addition to the same part of the krill trawl, also uses part of the 800 mm fore net, is shown in Appendix 8.

The trawl was rigged in exactly the same way as the 800 mm meshsize trawl, i.e., with 100 m bridles and front weights of 230 kg each. In this configuration (tows 22 to 30) at a towing speed of about 2.1 knots and a warp length of 250 m, the opening height was in the order of 26 m and the opening width of 30 m, giving an area of about 780 m², which is very close to the calculated values given in the table in Section 3.

In order to enable still shorter warps without loss in opening height, when fishing for D1 layers at low speed, for tow 32 and following (Fig. 7), the front weights were increased to about 360 kg each. This proved to be not sufficient for the higher speeds required for fishing with 200 m warps only in depths of 80 to 120 m. Consequently, the front weights were further increased to the original weight of 520 kg each at tow 38. This rig was maintained until tow 51, after which the front weights were again reduced to 280 kg each to facilitate night trawling with light at about 50 m depth and with only 100 to 150 m warps.

Appendix 4 shows that the variations in opening shape and size were larger for the 1600 mm than for the 800 mm trawl. This is due to the variations in rig just mentioned and also to a larger variety in fishing pattern. Because of partly shorter warps and/or higher towing speed the average opening size of the 1600 mm trawl was comparatively smaller than that of the 800 mm trawl. This applies particularly for night trawling with light when the opening area fell far short of the figures given in the table in Section 3. There is considerable scope for gear improvements (mainly rigging) for near the surface trawling through future trials.

Otherwise, this trawl performed and could be handled in the same satisfactory way as the 800 mm trawl, only that shooting and hauling took a bit longer because of the greater net length.

Of particular interest is the catching efficiency on reasonably good D1 layers, which is almost exactly the same as was found for the 800 mm trawl. This would indicate that also the meshes of double the size had still an appreciable guiding effect on the small fish.

Of greatest significance regarding the prospects of a future commercial midwater trawl fishery is the outcome of the last tow (58). This was conducted by Capt. R. Drange while the consultant was taking a well earned nap after more than 28 hours of continuous fishing. During searching along the edge of the continental shelf on the way back to Muscat, a D1 layer was finally found of a density which reminded the Norwegian scientists of conditions found during previous surveys.

Judging from the integration deflection, the fish density was in the order of 86 g/m³. Appendix 4 shows that this was almost four times more than the best density (tow 20, 24.2 g/m³) and many times more than the average for the rest of the trials. Although this about one half hour tow was not well recorded, the very impressive catch of about 50 t, of which 13 t could be hauled on deck after releasing the excess, proves that with similar or even only half this fish abundance, commercial fishing would be possible with the trial gear (See Fig. 8, 9 and 10). It should be mentioned that the catching efficiency was in the same order as for much lower fish density. Future trials will have to show whether such high fish density is more common than experienced during these trials or whether it is a rare exception. Indications from former surveys are rather encouraging.

The net damage caused by this large catch is, naturally, avoidable by stronger net construction. The handling of catches of this size suggests the use of a fish pump.

5.4 Trawling with Light

The experiments with artificial light conducted by the Norwegian scientists confirmed earlier observations according to which B. pterotum is certainly not attracted, but rather repelled. This is not surprising, because this species descends at daybreak to a depth of 100 m and more and ascends to near the surface only around sunset. More details of these experiments will be reported by the cruise leader.

The conclusion this consultant drew from these observations was that the repulsion by artificial light, which leads to a reversal of fish concentrations in the otherwise rather uniform N1 layer towards the lower edge, might be utilized for night trawling. Sample trawling with the Harstad trawl in the N1 layer without light had not given any encouraging results.

A first fishing trial with the 1600 mm trawl (tow 47) was immediately successful by yielding 5.6 t in 1 h. Consequently, a 24 h simulated commercial trawling test was started. No time was spent on searching but towing was back and forth, taking the traces as they came. The total catch (tows 47 to 54) was 40 t, with an average of 2.7 t/h. This included 15.8 t from 7 h of night trawling with artificial light, with an average catch of 2.3 t/h. This is not overwhelming but still quite promising, taking into account that the average net opening area was only 430 m² and could be increased considerably.

Also, the lighting was improvised, adding only a few overboard lamps to the normal ship’s illumination, which falls predominantly on deck. There is no doubt that lighting could be considerably improved by providing more overboard lamps along the rail and particularly the stem and by providing extra searchlights to cover the propeller wake.

The net was towed, depending on the density in the ship’s sounders echotraces, with the headline in depths between about 30 and 60 m and it was the impression that the fish went even further down after the vessel had passed. This may be partly due to the propeller wake which, incidentally, did not seem to interfere with one-boat midwater trawling in more than about 30 m depth.

The utilization of fish repulsion by light for night midwater trawling is not a standard commercial fishing practice so far. The consultant has heard about German trawler skippers using the searchlight on the propeller wake to push the fish down from the surface when midwater trawling for herring during the night off the south Norwegian coast in the sixties.

For night midwater trawling on B. pterotum, the development of this technique could very well become a very significant contribution towards commercial operations. This should be one of the major targets of the proposed pilot project.