Monte Carlo Simulations & Sensitivity Analyses for Assessing & Optimizing Design of Integrated Multi-Trophic Aquaculture Sites
An aquaculture sustainability project in the Bay of Fundy, Canada, is successfully making the transition from an experimental to a commercial scale. Mussels and kelps grown beside salmon cages have demonstrated accelerated growth rates due to the augmentation of their natural food sources by nutrients generated from the salmon. This aquaculture practice, were the by-products of one species becomes the nutrient inputs for another is known as Integrated Multi-Trophic Aquaculture (IMTA). If properly implemented, the benefits of IMTA are twofold. Economic diversification is fostered by the culture of additional harvestable commodities within the same site licence area, and the overall nutrient load to the environment is reduced.
Several challenges however, need to be overcome for open-water IMTA to optimize sustainability. It is the ratio of nutrient releasing fed biomass (i.e. fish) to the nutrient converting biomass of co-cultured extractive species (e.g. mussels, kelps) in their respective biomitigating niches that largely influence nutrient recovery efficiency; not necessarily the physical/spatial scale of any one component. Consequently, rearing nutrient extractive species at scales complementary to the fed species presents novel challenges. ‘Trial and error’ learning approaches are largely unavoidable, due mainly to new husbandry and site design. Each species within the system also has unique temporal and spatial culture requirements, adding further complexity. Continuous site evolution and unpredictable dynamics are typical of commercial operations and present unique challenges to modeling the system.
Nevertheless, modeling approaches, like Monte Carlo simulation, can generate a likelihood of outcomes based on ‘partial data’ thereby providing practical estimates until validation can occur at ‘fully evolved’ commercial sites. The use of @RISK software combined with nutritional mass-balance models in Excel has been ideal for this approach; simplifying otherwise complex processes and ‘reporting’ results in a manner understandable to all stakeholders. The resultant sensitivity analyses are providing strategic research and management direction by identifying variables that most affect the system.