Introduction
Polyculture, defined as the simultaneous rearing of multiple fish species within the same system, tends to be most efficient when co-occurring species exploit distinct trophic resources. A key challenge is therefore to minimize trophic competition, which can otherwise reduce growth through detrimental interactions and suboptimal resource use. The intensity of trophic competition can be estimated from the trophic distance between species, based on trophic functional traits reported in the literature (Lecocq et al., 2024). However, distances based on trophic trait proxies have not yet been compared to realized trophic interactions under field conditions. The reliability of the trophic trait approach therefore remains to be established using methods capable of quantifying such interactions in experimental conditions such as stable isotope analysis.
The objective of this study was therefore to assess whether trophic distances derived from species trophic functional traits are consistent with those obtained from stable isotope mixing models (SIMM), to determine the reliability of the functional trait-based approach as a proxy for trophic competition in polyculture systems.
Method
The assessment was conducted in two ponds containing both stocked and wild fish species, alongside cage-fed tilapia. Distances were then computed using standardized matrices. For the functional trait approach, the matrix reflected the relative contributions of different trophic functional groups selected from the FishBase classification. For the stable isotope approach, fish muscle and multiple potential food sources (aquatic macrophytes, insects, crustaceans, snails, fish, soil and suspended organic matter, biofilm, and commercial pellets for tilapia) were sampled in both ponds and analyzed for δ13C and δ15N to estimate source contributions to fish species diet. The standardized matrix was subsequently constructed from the normalized mean contributions of these sources, as inferred using the SIMM. The Bray–Curtis distance was used as the metric for both approaches. The performance of the functional trait approach was assessed by comparing the distance matrix derived from trophic traits-based approach with that obtained from the SIMM, using a Mantel test.
Results and discussion
In both ponds, significant positive correlations (r2 > 0.5; p-value < 0.005) were detected between distances derived from the trophic trait-based approach and those obtained from the SIMM. These results suggest that the functional trait-based method can be accurate when species diets are well documented. However, substantial discrepancies occurred for some species, notably due to pellet consumption from cages and context-dependent feeding behavior which can be influenced by resource availability, density, and individual preferences. Overall, we showed that this functional trait-based approach offers a rapid, cost-effective and practical framework for selecting fish assemblages with minimal trophic competition to improve polyculture design. Based on this, we further illustrated an application of the trophic trait-based approach to rank combinations drawn from a larger candidate pool within the two study ponds.
References
Lecocq, T., Amoussou, N., Aubin, J., Butruille, G., Liarte, S., Pasquet, A., & Thomas, M. (2024). Stronger together: A workflow to design new fish polycultures. Reviews in Aquaculture, 16(3), 1374-1394. https://doi.org/10.1111/raq.12903