Trawling is a relatively non-selective method of fishing, resulting in the discard of non- commercially important species that are captured as bycatch. In December 1998, the Australian Standing Committee on Fisheries & Agriculture finalised the National Policy on Fisheries Bycatch (SCFA 1998). The Policy was developed to provide a national framework for coordinating action to address bycatch issues and in June 1999 the WA government adopted this national policy as its own. As part of this policy, the Department commenced trialling and the implementation of bycatch reduction devices (BRDs) into all WA prawn and scallop trawl fisheries in 2001 (Bunting 2002).
BRDs fall into two categories: primary bycatch reduction devices (i.e. grids) are those that physically exclude large organisms allowing them to pass out of the net; and secondary bycatch reduction devices, such as square mesh panels (‘fish exclusion devices’ [FEDs]), are more passive devices that take into account the behavioural differences between target and bycatch species in order to allow bycatch species to escape (Broadhurst et al. 2002).
In 1998/99, experimental trials of grids were undertaken in WA using grid types used in other Australian trawl fisheries and in the United States (Watson & Taylor 1996; Robins &
McGilvray 1999; Olsen 1999). A few fishers also trialled several grids independently, indicating a proactive approach by some operators to reduce bycatch before any government legislation was considered. However, subsequent adoption by industry of some of the grid types trialled during this experimental phase showed that these grids were not effective in eliminating large animals and / or bycatch without substantial loss of target species under some conditions.
Hence, an FRDC-funded project (i.e. Kangas & Thomson 2004) was initiated to tailor BRD usage to the specific requirements of a number of WA trawl fisheries, including the SBPMF, the SBSMF and the Exmouth Gulf Prawn Managed Fishery. As the SBPMF and SBSMF operate within the Shark Bay World Heritage Property, part of the project was to identify those species important to the overall values of the World Heritage Property and develop gear that was successful at reducing bycatch of those species in particular (Kangas & Thomson 2004).
Although there are some similarities in fish species between Shark Bay and Exmouth Gulf, there are differences in bottom type and bycatch species affecting the efficiencies of BRDs and their impact in the fishery. For example, due to the large seagrass beds in Shark Bay, the trawl fishery has problems with capturing large amounts of free-floating ‘weeds’. Fishers in the SBPMF must also consider scallop bycatch during periods when the scallop season is closed. In order to tailor BRDs to the different fisheries, key fleet personnel travelled to Queensland and met with people experienced in the design, construction and use of BRDs and used the information from these fishers to decide which BRDs to trial in WA (Kangas &
Thomson 2004).
Commercial catch and bycatch information is available from grid trials in Shark Bay in 2000 and 2001. Departmental observers were used to record commercial catch and bycatch for most trawl shots conducted, with a ‘shot’ defined as a trawl of two nets, for between 30 minutes and three hours. During the trials, each boat typically towed two types of nets, a control or standard net and one fitted with some type of BRD. Eight main types of grids were trialled in Shark Bay including various combinations of:
• circular or rectangular;
• straight vertical bars or a horizontal ‘flounder’ gap at the bottom;
• narrow or wide bar spacing; and
• If top section of grids had bars that are bent as an accelerator or not (Kangas &
Thomson 2004).
The categories recorded for each side were: total bycatch weight or volume (including small or juvenile fish, crustaceans, echinoderms and molluscs); target species catch and component that is soft and broken (king, tiger, endeavour, coral prawns and scallops); and numbers of sharks, rays, sea snakes, sponges and turtles (i.e. ETP species). Pink snapper and tailor were also noted in Shark Bay due to their commercial and recreational importance (Kangas &
Thomson 2004).
During the grid trials, 1180 trawl shots comprising 1237 hours of trawling were recorded by observers on board commercial vessels in Shark Bay. Grid type was found to be a significant factor in total prawn catches, but did not result in a significant increase in the proportion of soft and broken prawns (Kangas & Thomson 2004).
Differences among grid types were significant on the amount of bycatch retained in nets
were 87 % less sharks and 88 % less rays retained by nets with grids compared to the control nets, with the majority of sharks caught in the control nets less than one metre in length. Only a marginal difference was observed with catches of ‘other large finfish’ (excluding pink snapper and tailor) between nets with grids and control nets; however, only a small number of trawl (n = 19) recorded catch of ‘other large finfish’. Overall, there was a 9 % reduction in scallop catch in nets with grids compared to the control nets, with no significant difference between grid-types (Kangas & Thomson 2004).
From a total of 914 trawl shots with a grid on one side and the control net on the other, 20 turtle captures were observed on the control side (no grid) and only one turtle capture was recorded with a BRD-net. This represents a 95 % reduction in the number of turtles caught.
The captured turtle was returned to the sea alive. The number of sea snakes caught in nets with grids was also 42 % lower than in control nets, which may be related to the movement of weed and snakes out of the escape opening (Kangas & Thomson 2004).
Grid type was found to have a significant impact on sponge catches, with the reduction of sponges retained in nets varying between 79 and 100 %. Differences were also observed between the proportion of weed retained by different grid types compared to control nets, with between 79 and 96 % reduction in most commonly used grid types (Kangas & Thomson 2004).
Two experimental trials of grids and secondary BRDs were also completed on established prawn-trawl grounds in Shark Bay in August 2000 using two chartered commercial prawn trawlers. The experiments were done using a twin-rig system (each with a headline length of 14.6 m), with all trawls made from polyethylene twine with a stretched mesh size of 52 mm in the body and 47 mm in the cod end. All tows were done over a combination of sandy and light coral bottoms in depths ranging from 13.7 to 18.5 m and at speeds of between 3.5 and 4.6 knots (Broadhurst et al. 2002). The grid trialled was an industry-designed aluminium grid with the upper third offset at 45° and bar spacing of 100 mm (Figure 9.1a). The grid was located at a 45° angle in a 30-mesh extension piece, with the anterior end attached to the trawl body. The BRD had no funnel or guiding panel but included two flexible panels of 47 mm diamond-shaped mesh hung loosely above and below the escape exit (Figure 9.1b).
Three cod ends were constructed and rigged with zippers so that they could be attached posterior to the extension containing the grid. The first cod end was a conventional 47 mm diamond-shaped mesh, while the two other cod ends were composite square mesh panels (aft [CSMPA] and forward [CSMPF]). These cod ends included secondary BRDs comprising composite panels made of 47 mm, 94 mm and 155 mm mesh cut on the bar and inserted into the top sections of the cod ends (see Kangas & Thomson (2004) for more detail).
Using zippers, the conventional (i.e. no BRD), CSMPA and CSMPF cod ends described above were alternatively attached posterior to the grid (Figure 9.2) and the entire assembly tested against the control cod end, on each side of the twin-rigged gear (i.e. three separate paired comparisons: grid only - i.e. no secondary BRD vs. control; grid and CSMPF cod end vs. control; and grid and CSMPA cod end vs. control). Two replicate 40-minute tows of each
paired comparison were made on each night, providing a total of 10 replicate comparisons of each configuration over five nights (Kangas & Thomson 2004).
Figure 9.1. Diagrammatic representation of (a) the grid used in Shark Bay experiments and (b) its location in the prawn trawl (Source: Kangas & Thomson 2004)
Figure 9.2. Diagrammatic representation of the prawn trawl with the Nordmøre-grid and location of the (a) conventional cod end, (b) composite square mesh panel aft (CSMPA) cod end and (c) composite square mesh panel forward (CSMPF) cod end (Source: Kangas & Thomson 2004)
Results from the experimental trials indicated differences in bycatch for the control nets, the grid only and the grid in combination with a secondary BRD. Compared to the control, the grid in combination with the CSMPA cod end significantly reduced the weight of bycatch (by 48.9 %) The grid only and the grid in combination with the CSMPF cod end showed some reduction in weight of bycatch, by 4.9 and 15.5 %, respectively (Figure 9.3). Trials on commercial boats indicate around 20 – 30 % reduction in overall bycatch, with up to 70 % reduction of some individual fish species (see Kangas & Thomson 2004 for species-specific reductions).
Figure 9.3. Differences in mean catch (± SE) of bycatch between the BRD and the control nets in Shark Bay. Black bars represent significant reductions. NG: Nordmøre grid;
CSMPF: composite square-mesh panel forward cod end; CSMPA: composite square-mesh panel aft cod end (Source: Kangas & Thomson 2004)