A Comparison between Surface and Underwater Feeding Technology with Respect to Feed Waste

Due to the technological development within the fish farming industry, new feeding technologies are being developed — however, little documentation concerning the sustainability of different innovations exist. This thesis explores different feeding technologies in aquaculture to compare their feed waste production with the use of a mathematical model of the feeding process.
Factors that are considered relevant during a feeding process are discussed together with challenges of feeding farmed fish. Two feeding technologies that currently are attempting to address these challenges are presented. First, the conventional surface rotor spreader is introduced, followed by a subsea feeder. The comparison is made possible by using an already existing mathematical pellet distribution model made by SINTEF Ocean for the surface rotor spreader. The model is altered to include the feed distribution pattern of a subsea feeder. It is of interest to see how the subsea feeder measures up to the surface rotor spreader. That is why two of the subsea design parameters, the feeding depth, and the bottom ring radius, are investigated. Both parameters are simulated three times with different values for three different water current profiles: uniform current, linear decreasing moderate current, and linear decreasing strong current. A behaviour related parameter, the local fish density threshold, FDthresh, is also simulated with two arbitrarily chosen values to show how important this parameter is to the overall model. To have a starting point for the simulation, a realistic feeding regime is developed and simulated in conjunction with operators at a fish farm.
The parameter study in this thesis reveals that the surface rotor spreader is the more sustainable option regarding feed waste, but it is uncertain if this is actually true in reality. For this model to produce more realistic results, it will be necessary for each fish farmer to implement local data from their fish farms. Further work must also be done to improve the model accuracy. The outcome of the comparison might change when the above is reviewed.