The presence of Atlantic bluefin tuna in northern European seas is driven by climate variability and ocean temperature.

Most demersal fishes undergo a dispersal phase as larvae, which strongly influences the connectivity among adult populations and, consequently, their genetic structure and replenishment opportunities. Because this phase is difficult to observe directly, it is frequently simulated through numerical models, most of which consider larvae as passive or only vertically migrating. However, in several locations, including the Mediterranean Sea, many species have been shown to swim fast and orient. Here we use a Lagrangian model to study connectivity patterns among three Mediterranean Marine Protected Areas (MPAs) and compare simulations in which virtual larvae are passive to simulations in which oriented swimming is implemented. The parameterization of behavior is based on observations for two groups of species of the family Sparidae: species with small larvae (i.e., 9–11 mm), displaying a maximum swimming speed of 6 cm s⁻¹ and a pelagic larval duration of 13–19 days (e.g., Diplodus annularis L., Oblada melanura L.) and species with large larvae (i.e., 14–16 mm), displaying a maximum swimming speed of 10 cm s⁻¹ and a PLD of 28–38 days (e.g., Spondyliosoma cantharus L.). Including larval behavior in the model (i) increased the overall proportion of successful settlers, (ii) enhanced self-recruitment within the MPAs, but also (iii) increased the intensity, and (iv) widened the export of eggs and larvae (recruitment subsidy) from the MPAs; overall, it significantly changed connectivity patterns. These results highlight the need to gather the observational data that are required to correctly parameterize connectivity models.

Through the advancement of observation systems, our vision has far extended its reach into the world of fishes, and how they interact with fishing gears—breaking through physical boundaries and visually adapting to challenging conditions in marine environments. As marine sciences step into the era of artificial intelligence (AI), deep learning models now provide tools for researchers to process a large amount of imagery data (i.e., image sequence, video) on fish behavior in a more time-efficient and cost-effective manner. The latest AI models to detect fish and categorize species are now reaching human-like accuracy. Nevertheless, robust tools to track fish movements in situ are under development and primarily focused on tropical species. Data to accurately interpret fish interactions with fishing gears is still lacking, especially for temperate fishes. At the same time, this is an essential step for selectivity studies to advance and integrate AI methods in assessing the effectiveness of modified gears. We here conduct a bibliometric analysis to review the recent advances and applications of AI in automated tools for fish tracking, classification, and behavior recognition, highlighting how they may ultimately help improve gear selectivity. We further show how transforming external stimuli that influence fish behavior, such as sensory cues and gears as background, into interpretable features that models learn to distinguish remains challenging. By presenting the recent advances in AI on fish behavior applied to fishing gear improvements (e.g., Long Short-Term Memory (LSTM), Generative Adversarial Network (GAN), coupled networks), we discuss the advances, potential and limits of AI to help meet the demands of fishing policies and sustainable goals, as scientists and developers continue to collaborate in building the database needed to train deep learning models.

Historically, the mortality of early‐life stages of marine fishes was supposed to be mostly caused by poor feeding during a critical period and aberrant drifting away from favorable recruitment areas. While fish larvae may display remarkable swimming abilities, Hjort's aberrant drift hypothesis has rarely been tested. In this study, we measured critical swimming speed (U_crit) of settlement‐stage larvae of six coastal, warm temperate Mediterranean fish species, for which no data were previously available (Sparidae: Boops boops, Diplodus annularis, Spicara smaris, Spondyliosoma cantharus; Pomacentridae: Chromis chromis; Mugilidae sp.). Their swimming speeds were comparable with those of other temperate species, but also with the speeds of tropical species, which are considered as very fast swimmers. Mugilidae were the fastest (29.2 cm s⁻¹), followed by Pomacentridae (22.8 cm s⁻¹) and Sparidae (11.6 cm s⁻¹). Most larvae swam in an inertial regime (Reynolds number > 1000). Those swimming speeds were then implemented in a Lagrangian model of the competency period of these species, set in the same area (the Ligurian Sea) and at the same time (June 2014) as the observations. In this modeling experiment, directional swimming strongly increased the proportion of successful settlers, independent of mesoscale hydrological structures. Fish larvae could settle on the coast from as far as tens of kilometers offshore, in just 4 d. These findings suggest that aberrant drift is unlikely to occur for strong swimming temperate larvae and show that larval behavior should be considered on equal footing with ocean currents when assessing larval fish dispersal.