Long-term variations in bycatch reduction device (BRD) effectiveness and trawl catch rates in a tropical fish assemblage

Fishing gears are often "non-selective". That is, they frequently catch organisms that are not the target of the fishery. Such organisms are usually referred to as bycatch. Some bycatch can be sold or used, but much of it is often of no value and unwanted. Thus, bycatch is a major problem in fisheries around the world, for both the fishery and the ecosystems from which the bycatch is removed. The prawn trawling industry has in the past captured very high ratios of bycatch to target species, due to the non-selective nature of trawl gear. Fifty to ninety percent of total catch by weight in a prawn trawl net can comprise unwanted species that are mostly discarded. Reducing the amount of unwanted catch in commercial fisheries is of major importance if we are to use fish stocks wisely and conserve global biodiversity. The implementation of bycatch reduction devices (BRDs) has been a major focus of fisheries management over the last two decades. The data for this study were obtained from field sampling in Cleveland Bay, North Queensland, Australia. The sampling was carried out by staff at James Cook University (JCU), assisted by students enrolled in a fisheries science subject at JCU. The equipment, sampling techniques and general area of sampling remained consistent over the entire fifteen year (1998-2012) study period. Prawn trawls were carried out for two or three days in the first two weeks of March for 12 of the 15 years, using the same vessel, the same net configurations, the same trawling procedures, and in the same inshore area of Cleveland Bay. Individual trawl times of 10 minutes duration could be expressed as a swept area of each trawl net of 6000m², and this also remained consistent over the entire study. For ease of comparison with the broader literature all catch rates are expressed as kg/hectare in this thesis. Thus, it was possible to obtain a long-term, scientific prawn trawling data set in which sampling procedures remained consistent. The first data chapter of this thesis examined the effects of the Jones-Davis BRD on catch rates of prawns and fish bycatch. A total of 244 trawls (each of 10 minutes duration) were carried out in 12 separate years between 1998 and 2012 by the JCU research vessel, James Kirby, rigged as a twin otter trawler. The two prawn trawling nets had different configurations: Turtle excluder device (TED) only (control net); TED and Jones-Davis BRD. The effect of the BRD on catch rates of four major teleost families, all other fish, penaeid prawns, and total fish catch was assessed. The BRD reduced catch rates of total fish bycatch, Carangidae, Sciaenidae and "Other Fish" significantly during the study. A significant year (time) effect on catch rates was detected for all variates investigated, but no significant interaction between the BRD and year of sampling was detected for any variates. Thus, the effect of the BRD was consistent across all years. Prawn retention using this device was high, with no significant difference in prawn catch rates between the two nets The BRD reduced catch rates of total fish (by 23.7%+/- SE 4.91), carangids (trevallies) (by 65.9%+/-SE 3.48) and sciaenids (croakers) (by 24.6% +/-SE 6.82) significantly. The BRD was an effective tool in reducing catch rates of fast, strong swimming semi-pelagic and pelagic fish species, but ineffective in reducing catch rates of slow swimming benthic and demersal fish species (e.g. Leiognathidae, Saurida spp.). Slow swimming benthic species comprise the majority of the tropical inshore fish assemblage. Averaged over the 15 years of the present study, the 23.7 % reduction in total fish catch means that the Jones-Davis BRD was ineffective in eliminating 76.3% of the fish bycatch in this tropical fish assemblage in Cleveland Bay. Attempts to design BRDs specifically for tropical Indo-West Pacific conditions have reduced fish bycatch in prawn trawl nets by 20-40%, at best. Thus, this study of the Jones-Davis BRD in Cleveland Bay is consistent with these other studies, even when the BRD was not designed for local conditions. The second data chapter of this thesis investigated the effect of environmental factors on inter-annual variations in trawl catch rates in a tropical fish assemblage in Cleveland Bay over 15 years. Environmental factors investigated included rainfall, tidal state and moon phase. Rainfall two years prior to sampling affected total fish catch rates significantly (p< 0.05; R² adj. = 0.305), again with catch rates enhanced by rainfall. The two year lag effect may be due to a two year lag in recruitment of fish to the fishery. This potential delayed recruitment into the fishery could be explained by migration of sub-adults out of the estuaries and mangrove forests into the near shore habitats where trawling occurred, combined with fish growing to a catchable size. Catch rates of prawns (mostly banana prawns, Fenneropenaeus merguinesis) were not affected by rainfall in this study. Catch rates of prawns and Saurida spp. (Lizardfish) were significantly higher on the full moon (p < 0.05). The effects of environmental drivers like rainfall and moon phase on inter-annual variations in catch rates of both prawns and fish in this tropical fish assemblage could only have been detected by consistent, long-term sampling. The third data chapter of this thesis examined inter-annual variations in the species composition of the tropical fish assemblage in Cleveland Bay from 1998 to 2012. A total of 160 teleost species were recorded in the 244 trawls made over 15 years. A Multi-dimensional scaling (MDS) analysis of the species presence-absence data demonstrated that there was no long-term, systematic shift in the species composition of the tropical fish assemblage in Cleveland Bay over the 15 years of study. That is, there was no evidence that the assemblage structure changed from one state to another, despite inter-annual variations in species composition and environmental conditions. In addition, there was little evidence of differences in species composition of the trawl catch between nets fitted with and without the Jones-Davis BRD. Long term monitoring studies are an important tool in fisheries research. Usually such studies involve collection of catch rate data from commercial fisheries. Gear efficiencies and fishing effort often vary in long-term studies of commercial fisheries. Long-term scientific, fishery-independent, surveys using completely consistent methods are rarer, often because funds to support them are limited. The long-term consistency of sampling in this study was critical in revealing inter-annual variations in performance of a BRD, identifying environmental drivers of trawl catch rates, and demonstrating the long-term consistency of species composition of a tropical fish assemblage