Search

Successful marine spatial planning relies on understanding patterns of human use, with accurate, detailed, and up-to-date information about the spatial distribution of fishing effort. In commercial vessels, tracking systems like the vessel monitoring system (VMS) or the automatic identification system (AIS) have helped to maintain and enhance the biodiversity of areas by generating large sources of positional data that served for commercial marine spatial planning. However, there is no regulation regarding location systems such as VMS or AIS for marine recreational fishing boats. Obtaining spatial data on marine recreational fishing can be difficult and time-intensive given the widespread and variable nature of the fleet. Remote cameras and computer vision systems are increasingly used to overcome the cost limitations of these conventional methods. Here we show a novel high-resolution and low-cost tracking system based on photo time-lapses and state-of-the-art computer vision algorithms, including deep learning, to automatically classify and obtain precise trajectories of fishing and cruising boats in coastal areas. Our method contributes to the automatic surveillance of marine protected areas by providing an image-based tool for automatic, real-time monitoring. Our method also allows for determining the intensity and spatial-temporal distribution of recreational fishing effort, important to defining the sustainability of the activity and coastal areas. We finally discuss the opportunities and limitations of computer vision tools applied to marine recreational fisheries spatial planning.

Chronotypes, the individual differences in daily activity timing, have profound associations with numerous physiological processes. Despite this, the covariance between chronotypes and other aspects of an individual's behaviour has been infrequently explored in non-human animals. This study delves into individual's variation across four axes of personality in a controlled environment, utilising the pearly razorfish, a model species for fish chronotype studies. We identified behavioural types across the aggressiveness continuum and established behavioural syndromes amongst exploration, activity, and boldness, irrespective of body size and condition. Subsequent to this, the experimental subjects were reintroduced to their natural habitat and individually tracked using high-resolution technology to ascertain their chronotypes. Our results revealed that whilst the exploration-activity-boldness syndrome bore no correlation with chronotypes, a significant association was observed between aggressiveness and chronotype. Hence, individuals with later awakening times and rest onsets were more aggressive than their counterparts with earlier awakening times and rest onsets. This study provides pioneering evidence linking fish chronotypes with other behavioural traits, such as aggressiveness, suggesting that behavioural variation could be potentially linked to the individuals' variation in internal clocks and the environmental variables influencing their expression.

Novel insights were provided by contrasting the composition of wild and farmed fish gut microbiomes because the latter had essentially different environmental conditions from those in the wild. This was reflected in the gut microbiome of the wild Sparus aurata and Xyrichtys novacula studied here, which showed highly diverse microbial community structures, dominated by Proteobacteria, mostly related to an aerobic or microaerophilic metabolism, but with some common shared major species, such as Ralstonia sp. On the other hand, farmed non-fasted S. aurata individuals had a microbial structure that mirrored the microbial composition of their food source, which was most likely anaerobic, since several members of the genus Lactobacillus, probably revived from the feed and enriched in the gut, dominated the communities. The most striking observation was that after a short fasting period (86 h), farmed gilthead seabream almost lost their whole gut microbiome, and the resident community associated with the mucosa had a very much reduced diversity that was highly dominated by a single potentially aerobic species Micrococcus sp., closely related to M. flavus. The results pointed to the fact that, at least for the juvenile S. aurata studied, most of the microbes in the gut were transient and highly dependent on the feed source, and that only after fasting for at least 2 days could the resident microbiome in the intestinal mucosa be determined. Since an important role of this transient microbiome in relation to fish metabolism could not be discarded, the methodological approach needs to be well designed in order not to bias the results. The results have important implications for fish gut studies that could explain the diversity and occasional contradictory results published in relation to the stability of marine fish gut microbiomes, and might provide important information for feed formulation in the aquaculture industry.

This dataset encompasses all necessary data required to replicate the study, `Chronotypes-Personality behavioural syndromes in wild fish’. The data were obtained through standardised behavioural tests conducted under laboratory conditions on 63 Pearly Razorfish (Xyrichtys novacula) individuals between April and July of 2019. Over a week, the fish were maintained in isolated aquariums to test their behaviours, including exploration, activity, boldness, and aggression, conducted daily. A Raspberry Pi system, equipped with the YOLOv5 deep-learning automatic tracking algorithm, was used to record these tests and calculate the fish's minute-by-minute position, providing essential data for evaluating exploration and activity. This system also stored videos to retrospectively obtain boldness and aggression data. Each test included only those individuals with at least two measurements. After the laboratory period, the fish were tagged with acoustic tags and returned to the sea to measure their chronotypes; only individuals with at least seven consecutive days of data were considered. The chronotype data, obtained from a previous study, are represented here through the previously derived scores. These laboratory-based experimental data were analysed using R software. In the exploration context, positional data were translated into total active time (TimeOut), minimum distance to the toy (MinDistance), and time spent near the toy (TimeToy). For activity, the data were converted into total active time (TimeOut), total distance covered (Distance), areas (CoreArea and Area), and direction angles (MeanAngle and KappaAngle). A Principal Component Analysis (PCA) was conducted to obtain the scores for exploration, activity, and aggressiveness. Upon acquiring these scores, trait repeatability was computed using a Linear Mixed-Effects Model, fitting the experimental day (Day), the total length of the individual (Size), and the internal condition (Condition) as fixed factors, and the i

Further investigation is needed to improve the identification and classification of fish in underwater images using artificial intelligence, specifically deep learning. Questions that need to be explored include the importance of using diverse backgrounds, the effect of (not) labeling small fish on precision, the number of images needed for successful classification, and whether they should be randomly selected. To address these questions, a new labeled dataset was created with over 18,400 recorded Mediterranean fish from 20 species from over 1,600 underwater images with different backgrounds. Two state-of-the-art object detectors/classifiers, YOLOv5m and Faster RCNN, were compared for the detection of the ‘fish’ category in different datasets. YOLOv5m performed better and was thus selected for classifying an increasing number of species in six combinations of labeled datasets varying in background types, balanced or unbalanced number of fishes per background, number of labeled fish, and quality of labeling. Results showed that i) it is cost-efficient to work with a reduced labeled set (a few hundred labeled objects per category) if images are carefully selected, ii) the usefulness of the trained model for classifying unseen datasets improves with the use of different backgrounds in the training dataset, and iii) avoiding training with low-quality labels (e.g., small relative size or incomplete silhouettes) yields better classification metrics. These results and dataset will help select and label images in the most effective way to improve the use of deep learning in studying underwater organisms

Fish differ consistently in behavior within the same species and population, reflecting distinct behavioral types (BTs). Comparing the behavior of wild and reared individuals provides an excellent opportunity to delve into the ecological and evolutionary consequences of BTs. In this work, we evaluated the behavioral variation of wild and reared juvenile gilthead seabreams, Sparus aurata, a highly relevant species for aquaculture and fisheries. We quantified behavioral variation along the five major axes of fish behavioral traits (exploration-avoidance, aggressiveness, sociability, shyness-boldness, and activity) using standardized behavioral tests and a deep learning tracking algorithm for behavioral annotation. Results revealed significant repeatability in all five behavior traits, suggesting high consistency of individual behavioral variation across the different axes in this species. We found reared fish to be more aggressive, social and active compared to their wild conspecifics. Reared individuals also presented less variance in their aggressiveness, lacking very aggressive and very tame individuals. Phenotypic correlation decomposition between behavioral types revealed two different behavioral syndromes: exploration-sociability and exploration-activity. Our work establishes the first baseline of repeatability scores in wild and reared gilthead seabreams, providing novel insight into the behavior of this important commercial species with implications for fisheries and aquaculture.

Deep learning allows us to automatize the acquisition of large amounts of behavioural animal data with applications for fisheries and aquaculture. In this work, we have trained an image-based deep learning algorithm, the Faster R-CNN (Faster region-based convolutional neural network), to automatically detect and track the gilthead seabream, Sparus aurata, to search for individual differences in behaviour. We collected videos using a novel Raspberry Pi high throughput recording system attached to individual experimental behavioural arenas. From the continuous recording during behavioural assays, we acquired and labelled a total of 14,000 images and used them, along with data augmentation techniques, to train the network. Then, we evaluated the performance of our network at different training levels, increasing the number of images and applying data augmentation. For every validation step, we processed more than 52,000 images, with and without the presence of the gilthead seabream, in normal and altered (i.e., after the introduction of a non-familiar object to test for explorative behaviour) behavioural arenas. The final and best version of the neural network, trained with all the images and with data augmentation, reached an accuracy of 92,79% ± 6.78% [89.24–96.34] of correct classification and 10.25 ± 61.59 pixels [6.59-13.91] of fish positioning error. Our recording system based on a Raspberry Pi and a trained convolutional neural network provides a valuable non-invasive tool to automatically track fish movements in experimental arenas and, using the trajectories obtained during behavioural tests, to assay behavioural types.

This data set is supplementary material for the article "Types of behaviour related to aggressiveness in pearlescent fish". Animal personality has a very relevant implication in a multitude of eco-evolutionary processes. It is a topic very approached in freshwater fish, not so in marine species, because its captivity is a challenge to reproduce its natural habitat. this is the first evidence of behavioural traits in the pearly razorfish. For this, we designed experiments in the laboratory, where the experimental individuals were subjected to the mirror test—a widely used test to determine each individual's aggressiveness. By not recognizing their reflection in the mirror, fish perceive their reflection as the intrusion of another individual into their territory. Thanks to these experiments, we were able to determine each individual's aggressiveness score, finding significant differences between individuals. Differences found regardless of height, sex or individual condition. These findings suggest that this species' aggressiveness may have a genetic origin, as has been demonstrated in other species.

Behavioural types (i.e., personalities or temperament) are defined as among individual differences in behavioural traits that are consistent over time and ecological contexts. Behavioural types are widespread in nature and play a relevant role in many ecological and evolutionary processes. In this work, we studied for the first time the consistency of individual aggressiveness in the pearly razorfish, Xyrichtys novacula, using a mirror test: a classic method to define aggressive behavioural types. The experiments were carried out in semi-natural behavioural arenas and monitored through a novel Raspberry Pi-based recording system. The experimental set up allowed us to obtain repeated measures of individual aggressivity scores during four consecutive days. The decomposition of the phenotypic variance revealed a significant repeatability score (R) of 0.57 [0.44–0.60], suggesting high predictability of individual behavioural variation and the existence of different behavioural types. Aggressive behavioural types emerged irrespective of body size, sex and the internal condition of the individual. Razorfishes are a ubiquitous group of fish species that occupy sedimentary habitats in most shallow waters of temperate and tropical seas. These species are known for forming strong social structures and playing a relevant role in ecosystem functioning. Therefore, our work provides novel insight into an individual behavioural component that may play a role in poorly known ecological and evolutionary processes occurring in this species.