Modelling of tuna around fish aggregating devices: The importance of ocean flow and prey

• Published in: Ecological modelling 2023-01, Vol.475, p.110188, Article 110188
• Main Authors: Nooteboom, Peter D., Scutt Phillips, Joe, Kehl, Christian, Nicol, Simon, van Sebille, Erik
• Format: Article
• Language: English
• Subjects: Fish aggregating devices; Individual-based model; Particle–particle interaction; Tuna
• ISSN: 0304-3800; 1872-7026
• DOI: 10.1016/j.ecolmodel.2022.110188

Catch and distribution of tuna in the ocean are typically investigated with ocean basin-scale models. Due to their large scale, such models must greatly simplify tuna behaviour occurring at a scale below ∼100 km, despite interactions at this level potentially being important to both catch and distribution of tuna. For example, the associative behaviour of tuna with man-made floating objects, that are deployed by fishers to improve their catch rates (Fish Aggregating Devices; FADs), are usually ignored or simplified. Here we present a model that can be used to investigate the influence of tuna dynamics below the ∼100 km scale on larger scales. It is an Individual-Based Model (IBM) of a hypothetical, tuna-like species, that includes their interactions with each other, free-floating FADs and prey. In this IBM, both tuna and FADs are represented by Lagrangian particles that are advected by an ocean flow field, with tuna also exhibiting active swimming based on internal states such as stomach fullness. We apply the IBM in multiple configurations of idealized flow and prey fields, alongside differing interaction strengths between agents. When tuna swimming behaviour is influenced equally by prey and FADs, we find that the model simulations compare well with observations at the ≲100 km scale. For instance, compared to observations, tuna particles have a similar stomach fullness when associated or non-associated to a FAD, tuna colonize at similar timescales at FADs after their deployment and tuna particles exhibit similar variations in continuous residence times. However, we find large differences in emergent dynamics such as residence and catch among different flow configurations, because the flow determines the time scale at which tuna encounter FADs. These findings are discussed in the context of directing future research, and an improved interpretation of tuna catch and other data for the sustainable management of these economically important species. •We present a model of tuna, and their distribution and catch below the 100 km scale.•Simulations resemble observed tuna colonization at FADs of ∼60–80 days.•Stomach fullness is higher for tuna particles when not associated to a FAD, similar to observations.•Both observed and modelled tuna residence times at FADs are often 1–10 days and sometimes tens of days.•Ocean flow has a relevant impact on tuna-FAD interactions.