Efforts to reduce or eliminate bycatch and discards in (trawl) fisheries have centred on gear modifications that increase the opportunities for undersized fish or unwanted species to pass through codend mesh or other selective device. Many gear modifications and bycatch reduction devices have proved effective in guiding and sorting fish (e.g. Watson and McVea, 1977; Andrew and Pepperell, 1992; Isaksen et al., 1992; Kennelly, 1995; Erickson et al., 1996; FAO, 1999; Rose, 1999; Broadhurst, 2000; Graham and Kynoch, 2001; Madsen, Holst and Foldager, 2002; Tschernij and Suuronen, 2002; Walsh et al., 2002; Valdemarsen and Suuronen, 2003; Broadhurst et al., 2004; Graham and Ferro, 2004; Graham et al., 2004; Sardà, Molí and Palomera, 2004; Valdemarsen, 2004; Fonseca et al., 2005; Polet, Delanghe and Verschoore, 2005). However, relatively little consideration has so far been given to concerns as to whether or not escaping fish survive.
It is generally assumed that effective selectivity automatically guarantees good survival. In many cases, this may be a fair assumption, but it should not be an automatic presumption. If survival is in doubt, it should be assessed. Improving selectivity without reducing the damage incurred by capture and escape is not an appropriate way of protecting immature fish. When developing and improving selectivity, it is important to address the whole range of stressors caused by the capture and selection process. This chapter presents some basic design principles that may help to reduce the injury and mortality associated with fishing and sorting processes.
It is well established that fish captured in a trawl often undergo forced swimming and experience contacts with the netting and other fish, crowding, confinement, crushing, barotrauma and oxygen depletion. Many groundfish often swim within or in front of the trawl mouth for a long time before they are overtaken by the gear (e.g. Main and Sangster, 1981; Wardle, 1993; Tschernij and Suuronen, 2002).
Little consideration has been given to the swimming capabilities of fish entering the trawl following this potentially exhausting process. When fish finally enter into the codend, they may be exhausted and therefore have greatly reduced swimming ability. Limited swimming capability may have implications on the efficiency of selection; that is, fish may be physiologically exhausted at the time they require a positive swimming action to escape (e.g. Breen et al., 2004). That is, not only may a fishs swimming ability determine its likelihood of capture by a towed fishing gear, but it may also influence the fishs escape and subsequent likelihood of survival. Moreover, the stress, exhaustion and potential injury to which fish are subjected when swimming in front of and inside the gear are probably cumulative. That is, the longer the capture process takes, the higher the likelihood of severe exhaustion and physical damage. The time between entering the fishing zone of the net and escaping should be minimized by proper gear construction and operational aspects. Greater consideration should be given to gear modifications that ensure that non-target fish are physically capable of making an active escape.
Fish should generally be allowed to escape before they enter into the codend, where the risk of damage is highest. This is not only because of crowding and crushing, but also because of problems caused by clogged meshes in the codend. Many investigations have demonstrated that codend meshes may become totally blocked by larger catches (e.g. Erickson et al., 1996; Suuronen, Erickson and Pikitch, 1997). In such a situation, escape from a codend is difficult or impossible. Erickson et al. (1996) showed by video observations that escape of walleye pollock occurred along the entire length (18 m) of the square mesh escape panel in front of the catch bulge. Hence, for this species, an escape panel or selective device installed ahead of the codend would effectively sort the catch by size before fish reach the codend meshes. It would also allow undersized fish to escape, even when the codend meshes become blocked with fish.
A similar approach has been suggested for many other fisheries, such as pelagic herring trawl fishery, where blocking of codend meshes is common (e.g. Suuronen and Millar, 1992; Suuronen, Erickson and Pikitch, 1997) and the fish species in question experiences high mortality when escaping from the codend (Suuronen et al., 1996b). These fish often lose scales easily, and swimming in the codend extension may cause severe damage to their skin. Fish may have lost a major part of their scales before they eventually reach the codend. For this type of fish species, escape panels and/or sorting devices should be installed as far forward as possible on the trawl, preferably in the trawl belly.
It is noteworthy that small (15 to 30 cm) gadoids (haddock, whiting, cod) excluded from a coastal shrimp trawl by a diagonal deflecting grid (Nordmøre grid) placed in front of the codend showed 100 percent survival (Soldal and Engås, 1997). Although some species can sustain stress and damage, it is always preferable to have selection in front of rear codends, where the likelihood of damage is highest.
Although there is little scientific evidence, it is intuitively obvious that mechanical sorting will cause more injury to fish than voluntary (i.e. active and oriented) escape from fishing gear. Voluntary escape requires that fish can detect the escape route. Vision plays an important role in the response and orientation of many fish species to trawls (e.g. Glass and Wardle, 1989), although there is evidence that some species are able to orient to the trawl under conditions of limited light (e.g. Engås and Ona, 1990). When light quantity is not adequate for visual mesh detection, fish generally have less chance of orienting properly towards a selective device (e.g. Olla, Davis and Schreck, 1997, Olla, Davis and Rose, 2000), and survival may be poorer (e.g. Suuronen et al., 1995).
Commercial fishing operations are often conducted at greater depths or at night, in dark conditions. Moreover, the codend is often surrounded by a dense mud cloud strirred by the ground gear. The design of trawls to enhance bycatch escape and survival must consider reactions under such conditions. There should be elements that will guide fish to the escape route. When some light is available at the fishing depth, the contrast of critical constructions should be maximized to facilitate the herding of fish into the right direction to increase escape. Contrast patterns between netting and the surrounding water can also be used in the manipulation of fish escape behaviour (e.g. Glass et al., 1995).
Management of the water flow inside the codend is an option when visibility is the limiting factor. Water flow can effectively guide at least smaller fish towards the selection panel or sorting device. Flow patterns can be managed by appropriate gear design and rigging, and through the use of various guiding and flow enhancement panels and other devices (e.g. Broadhurst, Kennelly and Eayrs, 1999; Engås et al., 1999).
Flow patterns that create a vortex within the codend or near any selection devices should be avoided because fish may rotate within the vortex for long periods; this may effectively prevent them from swimming and orienting properly towards the open mesh or selective device. Clearly, flow management within trawl codends may be highly important for selectivity and survival. It is worth noting that water velocity within a trawl gear generally decreases towards the codend (e.g. Thiele et al., 1997).
Operational aspects should be considered in detail when designing fishing gears that do not cause additional injury. The duration of a trawl tow and the catch size are issues that deserve particular consideration. Shorter tows with smaller catch sizes would help to maintain the selectivity properties of the codend, at least in trawl fisheries where catch rates are excessively high. On the other hand, it is notable that at very low catch sizes the selectivity of a codend may not be adequate because its meshes do not start to open properly until after the catch has built up. In demersal trawling, selectivity may gradually increase until a point is reached where it either levels out or begins to decrease as the catch quantity increases (e.g. Dahm et al., 2002).
It is clear that excluding debris and large animals and objects (such as boulders) will reduce the physical damage to fish caught in the codend, thereby increasing the survival chances of those that escape. By using improved ground gear constructions, sea bed objects such as boulders can be blocked. In addition, unwanted objects (e.g. rocks) that enter the trawl can be removed through various types of exit devices and holes. Such devices should be robust, effective, inexpensive and easy to manufacture, maintain and repair. It is also evident that knotless, smooth-surfaced netting materials should be favoured over stiff, knotted nettings in gear areas where the netting is causing abrasion of escapees (Figure 16).
Many gear-related factors may have a measurable effect on gear selectivity and the survival of escapees. These factors include mesh size and opening, number of meshes around the codend, length of codend and codend extension, length of lastridge ropes, thickness and stiffness of twine, construction of the codend lifting bag, attachments such as chafers, protection nettings and restrictor ropes, and overall gear rigging (see page 13). Modification of these structures can enhance escape and subsequent survival. However, there are limitations regarding the extent to which these structures can be modified, and much effort has recently been given to the development of various types of selectivity panels and bycatch reduction devices (BRDs).
Square mesh panels (windows) have great potential for improving trawl selectivity (e.g. Broadhurst, Kennelly and Gray, 2002; Fonteyne and Polet, 2002; Madsen, Holst and Foldager, 2002; Tschernij and Suuronen, 2002). When using square mesh panels, it is important to install them in the proper location on the trawl. Slight variations in panel design and location can dramatically affect selectivity and survival.
Even a very small panel may work efficiently when located in a strategically correct position. In the case of Baltic cod, the most efficient selectivity is likely attained when the square mesh panel is located in the aft most part of the codend (Figure 17). This is mainly because Baltic cod do not attempt to escape before entering the rear codend. However, it is also partly because the codend catches of cod are usually very small; if the panel were in front of the codend, the few fish swimming inside the rearmost part of the codend would not be able to see the panel. However, for most other fish species, it would be most beneficial to install the sorting panel or device in front of the codend, because many species incur most injuries while swimming inside the codend. This may require additional constructional elements that guide fish into the right position.
Selectivity and the survival of escapees may also depend on the materials used to construct the escape panels. The use of new materials such as Ulta-cross, Dyneema and composites may enhance selectivity and survival. Broadhurst, Kennelly and Gray (2002) suggest that owing to its flexibility, a composite square mesh panel fits the geometry of the codend better than square mesh panels made of plastic or metal do.
In many mixed species fishery, different species have to first be separated from each other (species selection), and can only then be sorted by size (e.g. Main and Sangster, 1982; Moth-Poulsen, 1994; Eayrs, Buxton and McDonald, 1997; Engås, Jørgensen and West, 1998; Broadhurst, 2000). Most selective devices that are intended to separate species operate by exploiting behavioural differences between different species. In order to be operative and effective, selective panels, guiding funnels and devices should be placed in strategic positions. In prawn trawls, for instance, these devices operate on the principle that fish have certain characteristic responses to towed trawls, whereas slower moving benthic invertebrates tend to show no specific, or very limited, responses (reviewed by Broadhurst, 2000). Underwater observations have shown that the majority of fish species attempt to escape in the upper part of the codend. This behaviour pattern is widely used in various bycatch reduction devices. Valdemarsen (2004) presented a new design principle in which deep water shrimp and fish are guided into separate codends by using a ring that separates the inner and outer codends. This principle is based on the observation that shrimp enter the trawl close to the bottom panels, while most fish pass into the codend at its centre. Such behavioural differences may be used to avoid fish bycatch in shrimp trawl fisheries; fish can be guided out of the trawl, and there is no need to use any filtering or sorting devices. Such mechanisms are likely to increase the survival chances of escaping fish (see e.g. Soldal and Engås, 1997).
Special guiding and sorting devices such as rigid and flexible grids can be used to effectively guide fish to an area where escape can occur, and species and sizes can be sorted (e.g. Isaksen et al., 1992; Rose, 1999; Kvalsvik et al., 2002; Fonseca et al., 2005). Various types of grids can be used as the components of complex systems. In species selectivity, grids work best when there are large size differences among the species that are to be separated. Behavioural aspects are critically important in effective grid sorting. Grid angle, change in angle, clogging of grid, handling properties, price and maintenance are important design and operational parameters (e.g. Isaksen et al., 1992; Broadhurst, 2000; Sardà, Molí and Palomera, 2004; Fonseca et al., 2005).
As a size-sorting device, a grid placed in front of the codend, in its intermediate or extension section may allow effective and consistent escape that is largely independent of catch volume (Figure 18). Effective grid sorting has been demonstrated for Atlantic cod, mackerel and Baltic herring when using longitudinal grids (e.g. Larsen and Isaksen, 1993; Suuronen, Lehtonen and Tschernij, 1993; Kvalsvik et al., 2002). One of the major advantages of grids is their stable performance under various conditions, particularly at higher catch rates, when conventional codend meshes tend to become blocked by the catch. However, blocking problems have also been encountered with sorting grids, and these should be taken into account in the design. There is some evidence that escape through rigid sorting grids may result in lower mortality compared with escape through a mesh (e.g. Suuronen et al., 1996b; Ingolfsson, Soldal and Huse, 2002; see also section on page 13). Clearly, the use of appropriate grid designs and installations may result in reduced escape mortality.
A fish may be run over by or collide with demersal trawl gear (e.g. Walsh and Hickey, 1993; Tschernij and Suuronen, 2002). Very little is known of the possible injury sustained by fish that are run over, but it could be speculated that some fish species may have a larger chance of survival when escaping between the rollers of a ground gear than when entering into a trawl and escaping later through a codend mesh. For some groundfish species this principle might be used as an alternative to codend selectivity (Figure 19). There are likely to be many more untested design options for developing selective ground gears in demersal trawl fishing. For instance, the rolling elements of the gear may be replaced with various types of weighted plastic sheet (Valdemarsen and Suuronen, 2003; Valdemarsen, 2004). Moreover, there are another ways of separating species and sizes based on behaviour. For instance, King et al. (2004) describe a flatfish trawl which has a cutback headrope designed to separate out rockfish prior to their entrainment inside the trawl. This type of trawl design permits nontarget species to escape before entering into the gear and has a significant potential to reduce bycatch mortality. Clearly, there is substantial scope for development in this area.
When attempting to reduce unaccounted mortality of young fish and non-target species, there are potential alternatives for the development of selective fishing gears. For instance, rather than developing techniques that enable the fish captured in a fishing gear to escape through codend meshes or other selective devices (e.g. grids), bycatches and their subsequent mortality could be reduced by preventing the fish from being captured in the fishing gear altogether. In other words, the capture of immature fish and nontarget species could be actively avoided by modifying the fishing strategy (e.g. Gauvin, Haflinger and Nerini, 1996; Witherell and Pautzke, 1997). In this way, fish would not be subjected to various stressors and physical injuries that result from capture and escape processes. The recent development of navigation and gear surveillance instruments should allow a marked improvement in the avoidance of hot-spots.
Although a great deal of progress has been made in reducing bycatch and discard through improving fishing gear selectivity, relatively little consideration has been given to the survival of escaping fish. Gear modifications to improve selectivity are generally based on the assumption that fish escaping from fishing gears are undamaged, minimally stressed and able to make a complete recovery. However, several studies have shown that physical damage incurred during capture and escape can result in fish mortality. Hence, evaluation of escape mortality should be an integral part of the development of selective fishing gears. Survival should ultimately be used to determine the suitability and success of new modifications.
Developing effective gear modifications that guarantee high chances of survival for escapees requires a good understanding of how fish react to gear under various conditions, including in situations when vision is limited or not operative. Clearly, immature fish that should escape from a fishing gear should stay inside the gear for as short a time as possible, and should not enter into the aft part of the codend where the risk of serious injury is highest. Installing escape panels or other sorting devices in front of the codend would probably enhance the escape and survival chances of undersized fish. It is evident that voluntary escape will cause less injury to fish than mechanical sorting. Hence, facilitating the voluntary escape of fish through appropriate constructions and operational improvements would increase the likelihood of survival. Use of non-abrasive netting materials, exclusion of debris and large objects from the codend, and use of better gear designs and riggings would further enhance the survival likelihood. It is clear that there is substantial scope for improving the survival of trawl escapees by using better gear modifications and operational solutions.
In general, measures to protect immature fish and non-target species are designed to increase the fishs opportunities to escape from fishing gears, rather than preventing or reducing their chance of encountering the fishing gear. Active avoidance of areas with high densities of juveniles and non-target species is an alternative approach to reducing unaccounted mortality.