A fully integrated GIS-based model of particulate waste distribution from marine fish-cage sites

• Published in: Aquaculture 2006-08, Vol.258 (1), p.299-311
• Main Authors: Corner, R.A., Brooker, A.J., Telfer, T.C., Ross, L.G.
• Format: Article
• Language: English
• Subjects: Agnatha. Pisces; Animal aquaculture; Animal productions; Aquaculture; Biological and medical sciences; Cage movement; Fish; fish cages; fish culture; fish farms; fish feeding; Fish waste dispersion; Fundamental and applied biological sciences. Psychology; General aspects; Geographic information systems; GIS modelling; mariculture; Marine; Marine biology; Marine ecology; marine fish; Marine fish cages; mathematical models; sediments; Vertebrates: general zoology, morphology, phylogeny, systematics, cytogenetics, geographical distribution; wastes
• ISSN: 0044-8486; 1873-5622
• DOI: 10.1016/j.aquaculture.2006.03.036

Modern Geographical Information System (GIS) offers a powerful modelling environment capable of handling large databases. It is a very suitable environment in which to develop a suite of tools designed for environmental management of aquaculture sites, including carrying capacity prediction, land–water interactions and multi-site effects. One such tool, presented here, is a fully integrated and validated particulate fish waste dispersion module which uses mass balance to estimate waste input and takes account of variable bathymetry and variable settling velocity for feed and faecal components. The model also incorporates the effect of cage movement on waste dispersion, the first such model to do so. When tidal range was low (1.67 m), the maximum movement of a 22 m diameter circular cage was 10.1 m and 7.7 m easting and northing, respectively. Highest deposition from particulate fish waste is under the cage and incorporation of cage movement increased the effective area under a cage by 72%. This reduced peak deposition measurements by up to 32% and reduced the average modelled feed and faecal settlement at the cage centre by 23% and 11%, respectively. The model was validated by comparing model predictions with observed deposition measured using sediment traps during three 2-week field trips at a fish farm on the west coast of Scotland. The mean ratio of observed to predicted waste deposition at 5–25 m from the cage centre ranged from 0.9 to 1.06, whilst under the cage the model over-predicts deposition (observed/predicted = 2.21). Although far-field data was seen to be comparable the near-field discrepancies resulted in variable overall accuracy in the model. The overall accuracy based on August 2001 data was ± 50.9%, on February 2002, ± 72.8% and on April 2002, ± 50.6%. Summarizing the data resulted in an overall average predictive accuracy of ± 58.1%.