|Geographical Information Systems (GIS) as a simple tool to aid modelling of particulate waste distribution at marine fish cage sites|
Pérez, O.M.; Telfer, T.C.; Beveridge, M.C.M.; Ross, L.G. (2002). Geographical Information Systems (GIS) as a simple tool to aid modelling of particulate waste distribution at marine fish cage sites. Est., Coast. and Shelf Sci. 54(4): 761-768
In: Estuarine, Coastal and Shelf Science. Academic Press: London; New York. ISSN 0272-7714, more
Cage culture; GIS; Modelling; Suspended particulate matter; Waste disposal; Marine
|Authors|| || Top |
- Pérez, O.M.
- Telfer, T.C.
- Beveridge, M.C.M.
- Ross, L.G.
Deposition of particulate organic waste from marine fish farm cages on to sea-bed sediments can cause major changes to the benthic ecosystem. Validated spatial models are considered as the most cost-effective tools for predicting environmental impacts. An improved version of an existing predictive particulate waste distribution model for farmed Atlantic salmon (Salmo salar L.) is presented, which uses Geographic Information Systems (GIS) combined with a spreadsheet. The model presented uses existing distribution algorithms but also incorporates functions to calculate feed loading for all the cages within a pontoon independently, spreads the input load over the whole cage area and simulates post-depositional distribution of the carbon. The model uses approximate estimates of feed and faecal waste derived from dietary considerations (mass balance model) and separate, unique settling velocities for waste feed and faecal particles. The model incorporates values of current speed and direction recorded over spring and neap tides. Output from the model is in the form of a contour plot of organic carbon (g C m -2), showing distribution of the particulate organic carbon material as deposited on the sea-bed. During this study using hydrographic data collected from near a fish farm, the model predicted a smooth gradient of sediment carbon concentrations which decreased with distance from the cages. Model performance was validated using measured levels of sediment carbon, and showed a significant correlation between predicted and actual sediment loading (R=0·7; P<0·01). The differences between predicted and measured quantities of carbon found at some sampling stations are likely to be due to processes not included in the model, such as small differences in bathymetry, differences in bottom type which may have increased or decreased the carbon distribution through saltation, or natural variation in the sediment composition.