Search Results (62)

Accurate and rapid aquatic organism recognition is a core technology for fisheries automation and aquatic organism statistical research. However, due to absorption and scattering effects, images of aquatic organisms often suffer from poor contrast and color distortion. Additionally, the clustering behavior of aquatic organisms often leads to occlusion, further complicating the identification task. This study proposes a lightweight object detection model, Mobile-YOLO, for the recognition of four representative aquatic organisms, namely holothurian, echinus, scallop, and starfish. Our model first utilizes the Mobile-Nano backbone network we proposed, which enhances feature perception while maintaining a lightweight design. Then, we propose a lightweight detection head, LDtect, which achieves a balance between lightweight structure and high accuracy. Additionally, we introduce Dysample (dynamic sampling) and HWD (Haar wavelet downsampling) modules, aiming to optimize the feature fusion structure and achieve lightweight goals by improving the processes of upsampling and downsampling. These modules also help compensate for the accuracy loss caused by the lightweight design of LDtect. Compared to the baseline model, our model reduces Params (parameters) by 32.2%, FLOPs (floating point operations) by 28.4%, and weights (model storage size) by 30.8%, while improving FPS (frames per second) by 95.2%. The improvement in mAP (mean average precision) can also lead to better accuracy in practical applications, such as marine species monitoring, conservation efforts, and biodiversity assessment. Furthermore, the model’s accuracy is enhanced, with the mAP increased by 1.6%, demonstrating the advanced nature of our approach. Compared with YOLO (You Only Look Once) series (YOLOv5-12), SSD (Single Shot MultiBox Detector), EfficientDet (Efficient Detection), RetinaNet, and RT-DETR (Real-Time Detection Transformer), our model achieves leading comprehensive performance in terms of both accuracy and lightweight design. The results indicate that our research provides technological support for precise and rapid aquatic organism recognition. Full article

Amid declining fish diversity and human pressures in freshwater ecosystems, robust basin-scale assessments are vital for effective fisheries management. This study collated nearly four decades of fishery yields from the Pearl and Yangtze Rivers to identify conservation priorities in the Pearl River Basin. It introduced a novel cumulative effect indicator based on zeta diversity—a biodiversity pattern metric—integrated with cumulative effects analysis for management decision-making. The research employed a multi-site generalized dissimilarity model to examine the non-linear relationships between fish species composition (ζ_n) and human pressures, environmental factors, and geospatial variations across elevation gradients. The cumulative effect indicator, reflecting responses to anthropogenic stress when assessing ζ₂ (related to β diversity), helped evaluate basins for conservation or restoration needs based on their unique or homogenized biotic communities. The results suggest that ζ diversity in low-elevation sub-basins has a stronger filtering effect on ζ by human pressures than in mid- to high-elevation sub-basins, where community aggregation is more random. The impact varied with diversity aspects (nestedness vs. turnover) and zeta order. A negative correlation between cumulative effects and community uniqueness validated the novel cumulative effect indicator’s effectiveness for guiding restoration in the Pearl River Delta, potential fishing bans, and karst conservation. This approach offers a theoretical basis for prioritizing areas for freshwater fish diversity conservation and fishing restrictions in the Pearl River Basin. Full article

The Mediterranean region is highly vulnerable to environmental and anthropogenic pressures, including climate change, which significantly affect its aquatic ecosystems, especially shallow lakes. This study examines the fish community and ecological quality of Lake Paralimni, a shallow mesotrophic lake in Central Greece that experienced complete desiccation between 1991 and 1996. Using field surveys, fish species composition, abundance, and biomass were assessed, and the lake’s ecological quality was evaluated through the Greek Lake Fish Index (GLFI) alongside an integrated SWOT (strengths, weaknesses, opportunities, threats) and AHP (Analytic Hierarchy Process) analysis. Six fish species from three families were recorded, predominantly native and endemic, with introduced species representing a minor fraction. While GLFI rated the lake’s quality as “Good,” other multi-metric indicators downgraded it to “Moderate”, highlighting the importance of comprehensive assessments. SWOT analysis revealed strengths such as high native biodiversity and legal protection under Natura 2000, but also weaknesses like fluctuating water levels and limited monitoring. Opportunities include sustainable fisheries and conservation efforts, while threats involve climate change, eutrophication, and illegal species introductions. AHP emphasized threats and weaknesses as top priorities. The study recommends hydrological regulation, invasive species control, and long-term monitoring for sustainable lake management and biodiversity conservation. Full article

Accurate fish population estimation is crucial for fisheries management, ecological monitoring, and aquaculture optimization. Traditional manual counting methods are labor-intensive and error-prone, while existing automated approaches struggle with occlusions, small-object detection, and identity switches. To address these challenges, this paper proposes an improved fish counting framework integrating YOLOv8-DT for detection and Byte-OCSORT for tracking. YOLOv8-DT incorporates the Deformable Large Kernel Attention Cross Stage Partial (DLKA CSP) module for adaptive receptive field adjustment and the Triple Detail Feature Infusion (TDFI) module for enhanced multi-scale feature fusion, improving small-object detection and occlusion robustness. Byte-OCSORT extends OC-SORT by integrating ByteTrack’s two-stage matching and a Class-Aware Cost Matrix (CCM), reducing ID switches and improving multi-species tracking stability. Experimental results on real-world underwater datasets demonstrate that YOLOv8-DT achieves a mAP₅₀ of 0.971 and mAP_50:95 of 0.742, while Byte-OCSORT reaches a MOTA of 72.3 and IDF1 of 69.4, significantly outperforming existing methods, confirming the effectiveness of the proposed framework for robust and accurate fish counting in complex aquatic environments. Full article

This paper presents a comprehensive time-series analysis framework leveraging the Temporal Fusion Transformer (TFT) architecture to address the challenge of multi-horizon forecasting in complex ecological systems, specifically focusing on global fishery resources. Using global fishery data spanning 70 years (1950–2020), enhanced with key climate indicators, we develop a methodology for predicting time-dependent patterns across three-year, five-year, and extended seven-year horizons. Our approach integrates static metadata with temporal features, including historical catch and climate data, through a specialized architecture incorporating variable selection networks, multi-head attention mechanisms, and bidirectional encoding layers. A comparative analysis demonstrates the TFT model’s robust performance against traditional methods (ARIMA), standard deep learning models (MLP, LSTM), and contemporary architectures (TCN, XGBoost). While competitive across different horizons, TFT excels in the 7-year forecast, achieving a mean absolute percentage error (MAPE) of 13.7%, outperforming the next best model (LSTM, 15.1%). Through a sensitivity analysis, we identify the optimal temporal granularity and historical context length for maximizing prediction accuracy. The variable selection component reveals differential weighting, with recent market observations (past 1-year catch: 31%) and climate signals (ONI index: 15%, SST anomaly: 10%) playing significant roles. A species-specific analysis uncovers variations in predictability patterns. Ablation experiments quantify the contributions of the architectural components. The proposed methodology offers practical applications for resource management and theoretical insights into modeling temporal dependencies in complex ecological data. Full article

According to European reports, the population of Anguilla anguilla has declined to unsafe biological limits in most areas and current fisheries are unsustainable. Indeed, the European eel has been listed as a critically endangered species since the 1970s and has been on the IUCN Red List of Threatened Species since 2010. Glass eel fisheries in Europe are very limited, but illegal catches and international parallel trade are major threats to eel stocks due to their high commercial value. The main hypothesis of this study is that glass eels from each estuary have unique chemical profiles according to the ecological quality of the habitat. These unique chemical fingerprints were assessed using Chemical Integrating Approaches (CIA) based on multi-element (macro, trace and ultra-trace metals), global metabolome and stable isotope analyses. Thus, CIA are intended to be an effective chemical “weapon” to (i) fingerprint wild glass eels; (ii) fingerprint captive glass eels; (iii) authenticate and trace glass eels; and (iv) combat the illegal trade in juvenile European eels. The results of this study showed that Zn and Ni can fingerprint the chemical profiles of wild and captive glass eels and their geographical origin. In the same vein, metabolomes such as Butyric acid 4-vinylphenyl ester, N-(3-carboxypropanoyl)-Met, 2-(4-Methylphenyl)acetamide, N-formyl-glutamic acid, 3-Hydroxy-2-(3-methylbutanoylamino)propanoic acid, 4-Dodecylbenzenesulfonic Acid, Arginine and Pyrazole and the stable isotope 15N show potential as a chemical tools for glass eel traceability. Full article

The inter-annual fluctuations of abundance of the Brazilian sardine (Sardinella brasiliensis) during the last decades have deeply modified the purse seine fishery dynamics. This study evaluates the trophic relationships of the main species exploited by this fishery and the importance of its biomass for the southeast–south Brazil marine ecosystem (22° S–34° S). Data were analyzed using a mass balance model (ECOPATH) between the two fishery collapses: 2000 and 2017. From 2000 onwards, the sardine fishery adopted a multi-species character. The mean trophic level of the catches (MTL) showed a decreasing trend until 2008, when more modern vessels with greater autonomy entered the fishery, and expanded the traditional fishing area to exploit northern fishing grounds. The MTL in the expanded fishing area suddenly increased and was characterized by high biomass of the Brazilian sardine and other species with a high biomass and high ecotrophic impact, falling again to the lowest level in 2016. The model evidenced high estimates for fishing mortality, natural mortality, and flow to detritus between 2008 and 2016, when sardine fishing collapsed. During this period, a sharp drop in the primary production required to sustain the catches from 2012 onwards accompanied a significant fall in the biomass accumulation rate. This pattern was coincidental with the increasing mean temperature of the catches, which was probably acting as a limiting factor for the primary production, resulting in a higher natural mortality and flow to detritus. Furthermore, the higher fishing mortality may have led the stock to collapse. Full article

In the context of ecosystem-based fisheries management (EBFM), multi-species models offer a potential alternative to traditional single-species models for managing key species, particularly in mixed-fishery settings. These models account for interactions between different species, providing a more holistic approach to fisheries compared to traditional single-species management. There is currently no comprehensive list or recent analysis of the diverse methods used to account for species interactions in fisheries worldwide. We conducted a systematic review to objectively present the current multi-species models used in fisheries. The systematic search identified 86 multi-species models, which were then evaluated to assess their similarities. Employing a clustering analysis, three distinct groups were identified: extensions of single-species/dynamic multi-species models, aggregated ecosystem models, and end-to-end/coupled and hybrid models. The first group was among the most diverse, owing to their ability to integrate biological components, while maintaining an intermediate level of complexity. The second group, primarily defined by the EwE method, features an aggregated biomass pool structure incorporating biological components and environmental effects. The third cluster featured the most complex models, which included a comprehensive representation of size and age structure, the ability to incorporate biological components and environmental effects, as well as spatial representation. The application of these methods is primarily concentrated on small pelagic and demersal species from North America and Europe. This analysis provides a comprehensive guide for stakeholders on the development and use of multi-species models, considering data constraints and regional contexts. Full article

Fisheries management requires improvement in scientific knowledge to ensure sustainable exploitation of important commercial species and population renewal. Within this context, this study focused on the reproductive biology of spot tail mantis shrimp, Squilla mantis, in the Central Mediterranean Sea, aiming to understand the reproductive pattern of this species and validate the macroscopic maturity scale through histological analysis. A multi-year sampling was performed from 2016 to 2020 by a commercial fishing fleet in the Northern Central Adriatic Sea (GSA 17), and a total of 2206 individuals were collected. The monthly average value of the total sex ratio of S. mantis was 0.524 ± 0.044 (mean ± SEM) in favor of females, which dominated the population from September to April. The proposed 5 stage macroscopic maturity scale was validated histologically through histological analysis, confirming synchronous ovarian development. The somatic indexes (GSI and K Fulton) and monthly incidence of macroscopic ovarian maturity stages highlighted a protracted reproductive season from winter to spring (January–May). Although the length-weight relationship showed a similar growth trend between genders, males reached a bigger size in terms of carapace length (C.L.) and dominated the population from 32 mm (C.L.). The macroscopic L50 estimated was 25.94 mm (C.L.). Full article

Target detection technology plays a crucial role in fishery ecological monitoring, fishery diversity research, and intelligent aquaculture. Deep learning, with its distinct advantages, provides significant convenience to the fishery industry. However, it still faces various challenges in practical applications, such as significant differences in image species and image blurring. To address these issues, this study proposes a multi-scale, multi-level, and multi-stage cross-domain feature fusion model. In order to train the model more effectively, a new data set called Fish52 (multi-scene fish data set, a data set containing 52 fish species) was constructed, on which the model achieved an mAP (mean average precision is a key measure of model performance) of 82.57%. Furthermore, we compared prevalent one-stage and two-stage detection methods on the Lahatan (single-scene fish data set) and Fish30 data set (a data set containing 30 fish species) and tested them on the F4k (Fish4Knowledge (F4K) is a data set focused on fish detection and identification) and FishNet data set (it is a data set containing 94,532 images from 17,357 aquatic species). The mAP of our proposed model on the Fish30, Lahatan, F4k, and FishNet data sets reaches 91.72%, 98.7%, 88.6%, and 81.5%, respectively, outperforming existing mainstream models. Comprehensive empirical analysis indicates that our model possesses a high generalization ability and reaches advanced performance levels. In this study, the depth of the model backbone is deepened, a novel neck structure is proposed, and a new module is embedded therein. To enhance the fusion ability of the model, a new attention mechanism module is introduced. In addition, in the adaptive decoupling detection head module, introducing classes with independent parameters and regression adapters reduces interaction between different tasks. The proposed model can better monitor fishery resources and enhance aquaculture efficiency. It not only provides an effective approach for fish detection but also has certain reference significance for the identification of similar targets in other environments and offers assistance for the construction of smart fisheries and digital fisheries. Full article

The application of deep-learning technology for marine fishery resource investigation is still in its infancy stage. In this study, we applied YOLOv5 and YOLOv8 methods to identify and segment fish in the seabed. Our results show that both methods could achieve superior performance in the segmentation task of the DeepFish dataset. We also expanded the labeling of specific fish species classification tags on the basis of the original semantic segmentation dataset of DeepFish and completed the multi-class instance segmentation task of fish based on the newly labeled tags. Based on the above two achievements, we propose a general and flexible self-iterative fish identification and segmentation standard workflow that can effectively improve the efficiency of fish surveys. Full article

The large variation in vertebral shape and calcification observed among elasmobranch species prevents using a single method for enhancing growth bands and reading age. Further, estimating age and growth parameters can be difficult or impractical when samples are incomplete due to the bycatch of a size-selective fishery. Using a single and rapid method, age readings were obtained on the vertebrae of four batoid species, namely 53 individuals of Dasyatis pastinaca, 51 of Raja asterias, 15 of Torpedo marmorata, and 55 specimens of Torpedo torpedo, from the local small-scale trammel net fishery in the coastal waters (5–20 m depth) of the Central Tyrrhenian Sea during 2019–2021. Based on these data, a statistical routine was developed to obtain multiple estimates of age and growth parameters for incomplete samples due to size-selective fishing. The acceptable agreement between and within readers (intra and inter-reader disagreement < 5%) and the rate of increase in vertebral size with body size (differently ranked across species) demonstrated the consistency of the enhancing method. The parameters estimated by the Von Bertalanffy and Gompertz growth models matched the data available in the Mediterranean Sea for the species studied, with D. pastinaca, T. torpedo, and R. asterias showing the lowest (k = 0.05–0.12), intermediate (k = 0.112–0.19), and highest (k = 0.18–0.23) growth rates, respectively, in line with the life history traits of these species. Overall, the method proved effective both in delineating band pairs in vertebrae of different species and in reliably estimating the age and growth parameters of problematic samples due to size-selective fishing. The proposed method supports the collection of comparable demographic data from other areas where similar multi-specific assemblages are bycatch of size-selective fisheries impacting potential nursery areas and other essential habitats for elasmobranchs. Full article

This paper presents the results of an analysis of stranding events of the Mediterranean monk seal Monachus monachus over a decade. The analysis involved categorization according to the cause of stranding and seasonality, the identification of hotspot stranding areas and an assessment of possible correlations between stranding events and environmental/climatic patterns using time series analysis. Moreover, Generalized Additive Models (GAMs) were applied to explore the effects of the size of small-scale fishing grounds, the number of species sightings, and the occurrence of reproduction sites on “human-related” strandings. Finally, special focus was put on the central part of the eastern Ionian Sea for the assessment of stranding hotspot areas by means of the Multi-Criteria Decision Analysis (MCDA) approach, based on different kinds of spatial information such as anthropogenic pressures and the location of breeding sites and feeding grounds. Time series analysis results revealed that oscillation indices, during the first half of the year, and sea surface temperature (SST) in the Mediterranean from October to December were positively correlated with monk seal stranding events. GAMs underlined that areas combining extended small-scale fishery grounds and a higher number of sightings were more likely to cause more strandings. Regarding spatial analyses, the central Aegean Sea was highlighted as a hotspot for “human-related strandings”, while the MCDA approach emphasized that the southern coasts of Cephalonia and the gulf between Lefkada and mainland Greece were susceptible to subadult strandings. Full article

The catch composition of a coastal gillnet fishery in the Eastern Mediterranean Sea was analyzed through a two-year experimental fishing survey. Seven fish species occurred regularly in the hauls. Surmullet, Mullus surmuletus, which is the most valuable demersal fish in Greek waters and the intended target of the gillnets in small-scale fisheries, was the most abundant and systematically caught species. Almost all surmullets were larger than their minimum conservation reference size. However, three commercially exploited species (Diplodus annularis, Pagellus acarne, and P. erythrinus) were caught systematically as undersized individuals. In addition, these three species were caught mostly as immature individuals. Moreover, the operational season of the surveyed métier overlapped completely with the reproductive period of five commercially exploited species (D. annularis, M. barbatus, M. surmuletus, Sphyraena sphyraena, and Trachurus trachurus). Improvements and the establishment of additional technical measurements should be considered for the small-scale gillnet fisheries in the studied area to attenuate their detrimental effects and achieve a better compromise between sustainable exploitation of the local multi-species fish resources and the need for an economically sustainable practice. Full article

The digitization of catch information for the promotion of sustainable fisheries is gaining momentum globally. However, the manual measurement of fundamental catch information, such as species identification, length measurement, and fish count, is highly inconvenient, thus intensifying the call for its automation. Recently, image recognition systems based on convolutional neural networks (CNNs) have been extensively studied across diverse fields. Nevertheless, the deployment of CNNs for identifying fish species is difficult owing to the intricate nature of managing a plethora of fish species, which fluctuate based on season and locale, in addition to the scarcity of public datasets encompassing large catches. To overcome this issue, we designed a transferable pre-trained CNN model specifically for identifying fish species, which can be easily reused in various fishing grounds. Utilizing an extensive fish species photographic database from a Japanese museum, we developed a transferable fish identification (TFI) model employing strategies such as multiple pre-training, learning rate scheduling, multi-task learning, and metric learning. We further introduced two application methods, namely transfer learning and output layer masking, for the TFI model, validating its efficacy through rigorous experiments. Full article