The development of fish production systems must be accompanied by a simultaneous reduction in their environmental impacts to ensure the sustainability of this sector. Fish polyculture emerges as one system that could meet these criteria, as several studies have found that it derives value from fish waste and increases nutrient recycling. However, it has historically been designed empirically, through trial and error, resulting in losses of time and money, thus limiting the exploration of diverse solutions. Therefore, this study aimed to develop a method to design a scenario that increases yield and optimises resource use and nutrient recycling. To this end, an iterative modelling approach based on trophic relationships was developed.
The trophic model scenario developed contained 10 trophic compartments, of which 9 were trophic groups with similar functional roles or diets, including fish (separating adults and juveniles) and their food resources available in the ponds: macroinvertebrates, zooplankton, phytoplankton, and detritus. The fish species were chosen based on the polyculture scenario of Aubin et al. (2021): common carp Cyprinus carpio (omnivore), pikeperch Sander lucioperca (carnivore), roach Rutilus rutilus (omnivore), and tench Tinca tinca (benthivore). Fish growth parameters and biomass and resource productivity were extracted from the same study to build the model. The scenario represented an experimental pond of 0.1 ha located in Le Rheu, France, (SEPURE project) with no distribution of formulated feed. By setting the stocking density at 87.5 kg/ha and modifying the percentage of each species in increments of 5%, 969 scenarios were simulated. R packages with mass-balance equations of Ecopath (Rpath) and ecological network analysis (enaR) were coupled to simulate and evaluate the scenarios. Each scenario was evaluated using performance indicators for productivity (net yield), resource-use efficiency (ecotrophic efficiency), and recycling (Finn cycling index).
The predicted net yield of the scenarios varied from 155-766 kg/ha, likely because the common carp had a higher specific growth rate than the other species. The ecotrophic efficiencies of phytoplankton (0.22-0.25) and detritus (0.41-0.52) varied little among scenarios. The ecotrophic efficiencies of juvenile fish, macroinvertebrates, and zooplankton exceeded 1 in 90% of the scenarios, which indicates that these compartments limited the productivity of the production system. The Finn cycling index clustered around 18%, 32%, or 42%. The seven scenarios that optimised these indicators (10-20% carp, 5-20% pikeperch, 35-50% roach, and 20-35% tench) seem the most interesting to explore. To improve selection of scenarios, it is important to know the potential natural productivity and characteristics of the ponds.