It is worth remembering that the collection of data suitable for detailed analysis was not an aim of these tests which had been planned first of all for demonstration and training.

Extreme caution must be exercised with the above reported results

The low number of trawl shots for each BRD, occasional damage to the codend and BRD, and variations in tow duration, location and direction means that the results are only suitable for preliminary analysis and interpretation. Furthermore, variation in fishing practices between the two fishing vessels makes it difficult to compare results.

When checking at the beginning of the tests that both the trawls were functioning correctly and measuring any catch bias between sides (with both port and starboard trawls equipped with standard diamond mesh codends), the number of baskets of shrimp were counted and both sides were found to be the same. Although this test was completed only once due to time constraints, it at least indicated that the new trawls were functioning correctly and that both sides were capable of catching the same amount of shrimp. It is worth mentioning that ideally this test should be repeated several times to reduce the likelihood that the catch result was due to chance alone. In this respect, it is worth drawing attention to the observations reported below, when testing Square Mesh Window which indicated that the trawls towed together were not performing as effectively on both sides.

When the tests first began the catch was categorised as either shrimp, commercial fish, large bycatch or small bycatch. Small bycatch included animals less than 15 cm in length. Much of the bycatch was approximately this length, however, time did not permit all bycatch to be measured and categorised as either large or small. All bycatch was therefore combined to form a single category. An additional problem with the catch categories occurred when small (juvenile) commercially important species were caught as these animals could be categorised as either commercial fish or small bycatch. To overcome this problem the category of commercial fish was maintained for those species normally too small to be retained for commercial purposes. This is not entirely satisfactory however as it misrepresents the catch of commercially viable fish and gives no indication of juvenile catch rates of these species. The same applies to categorising juveniles as bycatch.

The results indicate that all BRDs such as Square Mesh Window (SMW), preferably with Cone, or grid, NAFTED type, effectively reduced bycatch (with the exception of the fisheye).

Results on board the FV Mehrdad indicate that the introduction of the cone in combination with the SMW had little effect on bycatch exclusion but substantially reduced shrimp catches. The reasons for this shrimp loss are unclear and are in contrast to Bushehr results for this BRD. During these latest tests the SMW + Cone combination was only tested on the starboard side. Interestingly, in nearly all cases shrimp and bycatch catch rates were less on this side, despite highly variable catch rates. This indicates that the starboard trawl was not performing as effectively as the port trawl.

The NAFTED effectively reduced bycatch, large species in particular (such as rays, sharks or jellyfish) while maintaining catches of shrimp. It also excluded large numbers of stingrays from the catch which otherwise would have posed a safety hazard to crew sorting the catch. Large numbers of stingrays in the catch delays shrimp processing times and adversely affects catch quality, particularly during the heat of the day. The NAFTED substantially reduced catch sorting times and the exclusion of these animals by the NAFTED is likely to result in increased shrimp quality (although not tested).

When using fisheye, the shrimp loss was found extremely high. Possible reasons for this include location too close to the accumulated catch in the codend and loss through the escape openings during shooting and hauling of the trawls.

Regarding fish and shrimp surging forward when the codend was on the surface (prior to hauling the codend on board) which was observed with the video, Australian experience has found that shrimp loss increases with retrieval time and that losses increase in adverse weather due to surging action of the sea. A preventive measure is for the vessel to accelerate for a short period (immediately after the otter boards reach the towing blocks) and flush the catch back into the codend. Shrimp loss was visually observed during hauling of the codend with high numbers of shrimp escaping from the square-mesh window. Surging of the catch may also account for shrimp loss from the fisheye and the high losses onboard dhows reported by Shilat during previous tests.

During testing of the SMW v SMW + Cone combination, much of the catch fouled in a pocket of netting around the wire hoop (used to increase the effectiveness of the cone). This resulted in substantial shrimp loss and was caused by a combination of factors including the design of the trawl, the diameter of the wire hoop and the size of the catch. The seam joining the trawl extension to the codend was constructed of 240 x 40 mm extension meshes to 180 x 30 mm codend meshes. The hoop measured 600 mm in diameter. The hoop restricted the large number of larger extension meshes, allowing these meshes to 'balloon' outwards into a large pocket of netting and foul the catch. When hauling the codend onboard, shrimp and fish caught in the pocket were able to fall forward towards the SMW and escape through the meshes. This hoop was replaced with a 900 mm diameter hoop and was successfully used for the remainder of the tests.

The port side trawl and square-mesh window (SMW) also suffered extensive damage during the initial stages of the tests. This was due in part to poor attachment to the trawl and the design of the SMW. When first attached to the trawl, the ratio of trawl meshes to bars (of the SMW) along the top and bottom of the SMW was uneven. Following repair to the trawl a new SMW was constructed and attached to the trawl using a 3 to 1 mesh to bar ratio along the bottom of the window and a 2 to 1 ratio along the top. The rationale for this being that mesh opening in the trawl decreases towards the codend and therefore must be allowed for when attaching large windows. An 8 mm diameter rope was attached to the perimeter of the window for added strength.

Initially, the SMW measured 75 bar lengths long and 45 bar lengths wide. The design of the trawl and location of the window meant that the SMW was not only attached to the top panel, but also extended into the side panels. As the mesh opening of the side panels is likely to be different to the top panel, the SMW was modified so that it was attached solely to the top panel. The modifications included a 1P4B taper to reduce the width of the window (near the codend) to 28 bars.