Aquaculture Europe 2023

September 18 - 21, 2023


Add To Calendar 19/09/2023 14:00:0019/09/2023 14:15:00Europe/ViennaAquaculture Europe 2023THE FINS (FARMING IN NATURAL SYSTEMS) FRAMEWORK FOR AQUACULTURE MANAGEMENTStolz 2The European Aquaculture Societywebmaster@aquaeas.orgfalseDD/MM/YYYYaaVZHLXMfzTRLzDrHmAi181982


Joao G. Ferreira1+, Jon Grant2, Ramón Filgueira3, Ian Gardner4, Gregor Reid5, Leah Lewis-McCrea5, Kiersten Watson5, Anne McKee5, Alexander van Oostenrijk1


1 Longline Environment Ltd., 63 St Mary Axe, London, EC3A 8AA, United Kingdom

2 Dalhousie University, 1355 Oxford St, PO Box 15000, Halifax, NS B3H 4R2, Canada

3 Marine Affairs Program, Life Sciences Centre, Dalhousie University, Halifax, NS, B3H 4J1, Canada

4 University of Prince Edward Island, 550 University Ave, Charlottetown, PE C1A 4P3, Canada

5 Centre for Marine Applied Research, 27 Parker street, COVE, Dartmouth, NS B2Y 4T5, Canada

+Corresponding author,



Marine areas in the coastal zone have a variety of uses including fisheries, aquaculture, tourism, recreation, and other activities, all dependent on ecosystem services. Marine spatial planning (MSP) seeks to optimize coastal resource use to minimize conflict and maximize sustainability. In practical application, most MSP has been a GIS exercise in which the boundaries of various activities are mapped and perhaps zoned in order to reduce conflict.

However, these multiple activities have an influence on the environment that goes beyond, and is often not indicated by, their simple location.  Moving beyond static GIS layers to a more realistic portrayal of influences and footprints in the ecosystem provides much greater potential for management and is necessary in order to address carrying capacity and sustainability.Among the activities amenable to such management, aquaculture has huge potential for dynamic planning since (a) farm boundaries are specifically defined, (b) stocking density and biomass of cultured organisms is tightly controlled, (c) there is abundant scope for management operations including site location, stocking, harvesting, and fallowing, (d)  environmental interactions such as waste dispersion can be modelled and measured.

This is particularly the case for areas such as North America and Western Europe, where the legislative and governance frameworks are both established and well-tested, but the core principles that underpin sustainable aquaculture management are key to the development of the industry across the whole world and are a fundamental part of the drive for food security and food safety in light of population expansion and climate change.

The main objectives of FINS are:

  1. To produce a software platform for practitioners (i.e. farmers and managers) that allows for dynamic review of alternatives for siting, stocking, and general husbandry practice, taking into account both the environmental consequences of aquaculture and the ways in which the environment conditions aquaculture;
  2. To include and model a range of key performance indicators (KPI), ranging from compliance with loading thresholds to pathogen management and eutrophication, addressing both near- and far-field effects;
  3. To test the scientific validity of the framework and the degree to which accuracy and usability can be jointly optimised in order to deliver a robust product for practitioners.


FINS is used by means of the application shown in Fig. 1. Liverpool Bay, in eastern Canada, is shown in this example, together with various layers such as water circulation and dispersal of ammonia from the cage grid, a medium to far-field effect that is often ignored but a key driver for water column eutrophication.

The following set of FINS capabilities is highlighted:

  1. Georeferenced display of bathymetry for any area of the world. In the example in Fig. 1, a high-resolution dataset is used; in other parts of the world; sources such as GEBCO might be required;
  2. Display and animation of current velocities obtained from mathematical models and/or measurements from e.g. ADCP moorings; the bathymetry and current velocities are key to drive other models in the framework that simulate specific aspects of the aquaculture activity;
  3. Positioning and editing of aquaculture structures, definition of stocked species, stocking status, stocking density; in the example shown and results presented, the cultures species is Atlantic salmon (Salmo salar);
  4. Multiple grids can be placed in any coastal area ( or in a lake) to examine the interactions at a wider scale;
  5. Application of models to simulate growth of finfish (salmon, trout, seabass, seabream, tilapia etc) and shellfish (oysters, mussels, clams etc) in order to determine farm production and environmental externalities;
  6. Definition and export of a zone of interest for application of specific models to determine deposition, diagenesis, shellfish food depletion, and other KPI;
  7. Import of outputs from such models for display and analysis by FINS users. Fig. 1. shows imported layers for loading of POC to the sediment and for dispersal of NH4+ by advection and diffusion to a broad area;

The ORGANIX model (Cubillo et al, 2016) was designed to simulate loading of POC to the sediment, using both Eulerian and Lagrangian approaches. In addition, ORGANIX simulates emission and dispersion of ammonia, oxygen demand, and for bivalve shellfish, chlorophyll uptake and its consequences for food depletion. 

The FINS framework has been applied to a number of ecosystems on the Canadian east coast, including Liverpool Bay, Whitehead Bay, Port Mouton, and Saddle Island. Hydrodynamic inputs were generated through the application of the FVCOM model and optimised for offline coupling, since one of the priorities for FINS is a very fast execution.

Results and Discussion

The results provided in this paper are divided into two parts: (i) the first provides an illustration of the outputs from the ORGANIX model; and (ii) the second provides details of validation of outputs by comparison to previously published work.

The results in Fig. 2 (left) show the loading of POC to the sediment for three conceptual salmon grids, using an Eulerian model. As expected, the deposition fields in this relatively shallow bay are approximately bounded by the grid area; each cage is 30 m in diameter and stocked with 50,000 fish at an initial weight of 80 g, grown for 500 days, resulting in a final individual weight of 5 kg, with an FCR of 1.13.

In the central pane of Fig. 2, the same conceptual grids are used to grow mussels (1,500,000 individuals per structure). The culture starts with 1 g (live weight) seed and the animals grow to 19 g over a 730 day period. Each individual will clear 7.4 m3 of water during this period and remove (net) 26.3 mg of chlorophyll. The figure shows that there are ammonia peaks to the west of the structures, and unlike the deposition profile, a clear interaction can be seen among the three grids, although the ammonia concentrations at peak are very low (0.18 mM).

The right pane shows ammonia concentrations resulting from a trout (Oncorhynchus mykiss) grid at Port Mouton—this farm is no longer active, but a comparison can be made between these results and other work to provide validation for the FINS framework.

It should be noted that in all three examples given, the loading varies in time as determined by the underlying growth model—in other words the FINS approach does not use averaging for emissions, for both finfish and shellfish.

The modelling framework was validated for several bays in eastern Canada, including Liverpool Bay, Port Mouton and Whitehead Bay. Some results of validation for ammonia emissions from a legacy trout site in Port Mouton are shown below.

The dynamics of dissolved nitrogen were simulated for a historic farm in Port Mouton that usually held 400,000 steelhead trout over a production cycle of 16 months. Fish were typically stocked in cages at 150 g in April and were harvested at 2 kg during July or August of the following year. During the growout period, the fish were fed to near satiety.

The pens usually occupied the top 8-10 m of the water column, which averages 10-12 m depth at the farming site. This aquaculture scenario was simulated using two approaches: 1) a fully-spatial dynamic hydrodynamic model in which the mass of dissolved nitrogen released was estimated using a nutrient loading model (Reid et al. 2017), and 2) FINS, which uses the residual currents for the study area and the AquaFish individual growth model to estimate the release of nitrogen. In both simulations, the dissolved nitrogen is considered a conservative tracer in which the only source of nitrogen comes from fish excretion, and the only sink is the tracer that is exchanged through the boundary.

The outcomes of both simulations (Filgueira et al. 2021 and Figure 2 – right pane, for the hydrodynamic and FINS approaches, respectively) differ slightly in the shape of the footprint. While both approaches predict the major concentration of dissolved nitrogen on the eastern side of the vicinity of the farm, FINS predicts a low-concentration plume that extends beyond this area towards the northeast of the bay. The different shapes could be a consequence of comparing the output of a dynamic hydrodynamic model (Filgueira et al. 2021) with the integrated temporal pattern generated using residual currents (FINS).

Despite the difference in shape, it is important to highlight that the maximum values were in both cases comparable, with maximum values between 9 and 10 µM. Further, the total area with a concentration over 3 µM was similar and, in both cases, restricted to the vicinity of the farm. The similar outcomes suggest that the management decisions that could stem from both approaches would be similar. In this particular case, they would indicate the maximum concentration of dissolved nitrogen is below the toxicity levels for seagrass.

Although not shown due to space constraints, we present also simulations of pathogen spread using particle tracking models, and results on the application and validation of diagenetic models driven by POC loads simulated in ORGANIX. The representation of these results is shown in the FINS application.

FINS is applied to eight coastal systems in the Canadian Maritimes, in order to support management, increase sustainability, help create jobs, and contribute to food security.


The authors wish to acknowledge funding from the Atlantic Fisheries Fund, Canada, and the Horizon Europe NovaFoodies project. We are very grateful to Marko Jusup for the development of the particle tracking model for deposition of POC and helpful discussions on the use of particle tracking for pathogen dispersal.


Cubillo, A.M., J.G. Ferreira, S.M.C. Robinson, C.M. Pearce, R.A. Corner, J. Johansen 2016. Role of deposit feeders in integrated multi-trophic aquaculture - a model analysis. Aquaculture, 453, 54-66.

Filgueira, R., Guyondet, T., Thupaki, P., Reid, G.K., Howarth, L.M., Grant, J., 2021. Inferring the potential for nitrogen toxicity on seagrass in the vicinity of an aquaculture site using mathematical models. Journal of Environmental Management 282:111921

Inglis, G.J., Hayden, B.J., Ross, A.H., 2000. An overview of factors affecting the carrying capacity of coastal embayments for mussel culture. NIWA Client Report CHC00/69, Christchurch, New Zealand.

McKindsey, C. W., Thetmeyer,  H., Landry,  T., Silvert,  W., 2006. Review of recent carrying capacity models for bivalve culture and recommendations for research and management. Aquaculture, 261(2):451-462.

Reid, G.K., Forster, I., Cross, S., Pace, S., Balfry, S., Dumas, A., 2017. Growth and diet digestibility of cultured sablefish: Implications for nutrient waste production and Integrated Multi-Trophic Aquaculture. Aquaculture 470:223-229.