ABSTRACT This study examined diel activity, swimming patterns and vertical distribution of hybrid grouper juveniles ( Epinephelus fuscoguttatus ♀ × Epinephelus lanceolatus ♂) under captive conditions, providing behavioural reference data for the development of intelligent monitoring systems in grouper aquaculture. A total of 30 juveniles (mean total length TL: 11.49 ± 0.80 cm; body weight BW: 27.70 ± 5.83 g) were monitored in a recirculating aquaculture system using continuous 24‐h video recordings across three replicates (10 juveniles per replicate). Video analysis was used to quantify the proportion of active fish, three swimming patterns—hovering (HOV), fast swimming (FS) and vertical ascending (AV)—and vertical distribution across bottom, middle and surface layers on a per‐minute basis. Hybrid groupers were predominantly inactive, with only 28.55% of fish active over a 24‐h cycle. Clear diel rhythmicity was observed, characterised by crepuscular–nocturnal behaviour, with the proportion of active fish peaking at dawn (0500–0600 h; 41.00%) and dusk (2000–2100 h; 38.72%), and with a significantly higher ( p < 0.05) proportion of active fish at night (35.30%) than during the day (21.80%). A distinct midday peak in the proportion of active fish (1200–1300 h; 41.22%) was observed, which has not been previously reported in groupers. This peak coincided with the scheduled feeding time during the acclimatisation period prior to the experiment and was accompanied by elevated HOV and AV, reflecting feeding‐related behaviour. It was therefore attributed to food‐anticipatory activity (FAA), explicitly demonstrating its confounding effect on baseline behavioural data. Over the 24‐h period, more than 50% of the fish (58.8%–93.5%) consistently occupied the bottom layer. This finding confirms their benthic behaviour, consistent with their sedentary, ambush‐predator ecology. Overall, these findings provide the first comprehensive diel behavioural dataset for hybrid grouper juveniles and establish a refined baseline that accounts for FAA effects, thereby supporting the development of more accurate machine‐learning‐based monitoring systems in aquaculture.
Au et al. (Fri,) studied this question.