Waste outputs and dispersion around marine fish cages and the implications for modelling

Aquaculture plays an increasingly important role in food production as the catches of wild fish stocks continue to decline on a global scale through overfishing. However, the rapid development of intensive cage aquaculture in particular, which requires high inputs of energy, food and capital, can result in adverse effects on the environment. While spatial distribution and sediment loading models for particulate wastes from marine fish cages have been under development for more than 10 years, the models still contain numerous assumptions that limit their usefulness. These include the use of very limited data for fish feed and faecal pellets sinking rates that take no account of food manufacturer, type or size or environmental conditions. The present study provides information on a range of pellet types for three of the most important European farmed fish species (i.e. Atlantic salmon Salmo salar L., sea bream Sparus aurata, sea bass Dicentrachus labrax) that may be readily incorporated into models. Such data, combined with validation of predictions through in situ field investigations is designed to help improve the accuracy and usefulness of solid waste dispersal models. The studies comprise four main sections, i.e., quantifying food and faecal pellet characteristics, examining nutrient leaching rates from uneaten food and fish faeces, determining resuspension characteristics of uneaten food, modelling of solid wastes dispersion and thus the development of environmental tools. Existing literature relating to environmental impacts of cage aquaculture is reviewed and the key factors highlighted. Two preliminary studies provided information on the influence of gravity acceleration on settling velocity determination and appropriate techniques for monitoring the rate of nutrient leaching from faecal wastes. Settling velocities of Atlantic salmon diets were significantly greater at 20 psu salinity than at 33 psu and significantly higher for most pellet types at 10°C than, at 20°C. Settling velocities for unsoaked salmon diets were found to increase with pellet size, from a mean of 5.6 cm s-1 for the smallest pellet (2 mm) to 13.9cm s-1 for the 10 mm standard (20 to 24% fat) pellets. Settling velocities of extruded diets for sea bream and sea bass diets ranged from 3.9 to 10.6 cm s-1, broadly similar to those for salmonid diets. Settling velocities of salmon pellets were not significantly affected by immersion time (0 - 15 min). Given the water depths at fish cage sites and the settling times involved, it is concluded that it is unnecessary to take account of changes in food pellet settling velocity as a result of immersion. Freshly net-collected salmon faecal pellets appeared to consist of fine solid material approximately the size of the formulated diets. The range of salmon faecal settling velocities was 3.7 to 6.2 cm s-1 (mean = 5.3 cm s-1) at 15°C and 33 psu. There are no significant differences in nutrient leaching of carbon and nitrogen from all six salmon diets after 20 min immersion in sea water. However, a rapid loss of faecal nutrients occurred 2.5 to 10 min after immersion in sea water. Total C and total N were found to leach by as much as 22% and 26%, respectively, after 5 min immersion during one sampling occasion. Experiments conducted in a large-scale flume tank showed the critical resuspension velocities of a range of commercial fish feeds were between 8.63 cm s-1 and 9.53 cm s-1. Above the critical resuspension speed, pellets moved by saltation, i.e. traveling along the sediment by rolling, sliding or hopping on the bed. The velocities of pellet resettlement ranged between 0.79 cm s-1 and 3.98 cm s-1 under the critical resuspension speeds. Field trials, involving the deployment of sedimentation traps, showed a general relationship between sedimentation of material and distance from cages, i.e. more sedimented material was associated with sampling sites closest to the cages. The spatial changes in sedimentation rates in the first trial were between 15.4 and 31.7 g DW m-2 d-1 at 30 m and 10 m stations, respectively. Values in the second trial (38.5-65.5 g DW m-2 d-1) were twice those in the first trial, but followed the same pattern. The model presented in this thesis is a combination of a spreadsheet model (Microsoft Excel 6.0) and Surfer plot program (Golden Software Ltd., ver. 6.04). Excel is used to prepare basic mathematical operations behind the model, including a mass balance submodel and use of a formula for calculating dispersion of uneaten food and faeces on the sea bed developed by Gowen et al. (1989). The operation of the waste dispersion model for marine cages takes into account the various settling characteristics of waste particles. It was verified with a set of in situ sedimentation data obtained from the field trial described above. Results described the waste dispersal around the vicinity of the cage farm. For the future, it is intended that further validation and optimisation of the model will be carried forward by a combination of both increasing user involvement and incorporation of data from comprehensive studies as these become available. Together, these will contribute to reducing and remedying the environmental impacts of future development.