Until recently, concerns about the sustainable exploitation of marine resources focused mainly on proper management of the fish stocks targeted by directed fisheries. During the last two decades, however, growing concerns have been raised about the impacts on the ecosystem in general (Gislason, 1994). Questions about how towed fishing gears such as trawls and dredges may affect benthic habitats and organisms have attracted a great deal of attention, and consequently a large number of investigations addressing this issue have been conducted over the last decade. The reasons for such concern are that complex benthic habitats provide shelter and refuge for juvenile fish, and benthic organisms comprise important food sources (direct or indirect) for demersal fish (Auster et al., 1996; Collie, Escanero and Valentine, 1997).
Concerns about the use of towed fishing gear were raised by fishers as early as the fourteenth century in the United Kingdom, and the sixteenth century in the Netherlands (de Groot, 1984). Complaints concerned the capture of juvenile fish and the detrimental effects on benthic life as a source of food for larger fish. The first scientific impact investigation was conducted on plaice fishing grounds in the North Sea in 1938 (Graham, 1955). Starting in the early 1970s, comprehensive projects on the impacts of towed gears have been carried out by research institutes in countries bordering the North Sea (see de Groot, 1984; Lindeboom and de Groot, 1998; Kaiser and de Groot, 2000). In addition to the North Sea region, impact studies have also concentrated on fishing grounds in the northwestern Atlantic (Grand Banks, Georges Bank), coastal Australia/New Zealand and, more sporadically, the Mediterranean, the Bering Sea/Gulf of Alaska and coastal areas of the United States.
Despite the growing number of studies on this issue, few if any general conclusions have yet been drawn on the responses of benthic communities to the impacts of trawling disturbances (Collie et al., 2000). There are three principal reasons for the lack of knowledge in this research field. First, the structure of benthic communities is complex and shows large temporal (both seasonal and annual) and spatial variations. Anthropogenic disturbances may therefore be difficult to demonstrate because they are masked by more dominant natural factors. Second, studies to assess bottom fishing impacts show great variability in terms of gear type and design, disturbance regime, bottom type (e.g. mud versus cobble), level of natural disturbance and benthic assemblages studied. Considerable differences in the types of response to disturbances are therefore to be expected. Third, these studies show large variations in the methodology used and the scientific approach to the problem. These factors reflect the complexity of conducting impact studies, and show that several requirements have to be met before realistic conclusions can be drawn.
Owing to their complexity and methodological deficiencies, individual impact studies should be interpreted with caution. This report presents an evaluation of the methodologies used in studies, and accordingly reviews current knowledge of the physical and biological impacts of otter trawls, beam trawls and scallop dredges on benthic habitats and communities.
Several publications review experimental impact studies to some extent (e.g. Jennings and Kaiser, 1998; Watling and Norse, 1998; Auster and Langton, 1999; Hall, 1999; Collie et al., 2000), but none has taken into account the important caveats that limitations and shortcomings in the methodologies applied in these studies call for. Thus, the results from impact studies should be interpreted with caution, and the conclusions that can be drawn are often limited by methodological deficiencies. This review presents a critical evaluation of the methodologies used in impact studies, and uses it to make conclusions about what lessons have been learned to date on how benthic communities are affected by towed-gear fishing activities. It shows that most studies are far from fulfilling the criteria of an "ideal study", which requires analysis over long time scales and comparisons between fished areas and untouched control sites that do not change in other aspects.
