Fish Counts Research Articles

Preliminary result of population structure of reef fishes in coral reef restoration sites in Tun Sakaran Marine Park, Semporna, Sabah

Restoration of coral reefs at the Tun Sakaran Marine Park (TSMP) started in 2009. Various methods are adopted, although the Coral Frame method is preferred mainly due to the low maintenance cost and its durability, and is currently still being used. However, since the beginning of its deployment, there has been little study on the effectiveness of the coral restoration project. Thus, this study was conducted to determine the population structure of indicator coral reef fish around the coral frames. Two study sites (Site 1: Bohey Dulang; Site 2: Mantabuan) with existing coral frames within the TSMP were selected. At each site, a baited remote underwater video system ( BRUVS), was deployed, and each was set to capture approximately one-hour footage. The first BRUVS deployment of this 24-month project was made on 21 July 2020, during the peak of the Southwest Monsoon. A total of 20 families (32 species) and 19 families (42 species) were recorded at Site 1 and Site 2, respectively. The Small-tooth whiptail, Pentapodus caninus (MaxN:14) and fusilier, Caesio sp. (MaxN: 101) are the most abundant species at Site 1 and Site 2, respectively. The preliminary findings reveal a low abundance and diversity counts of indicator reef fish (Serranidae, Labridae, Scaridae and Lutjanidae) at the coral restoration sites within TSMP. However, it is suspected that opportunistic local fishermen try their luck to fish in restricted parts of the park (pers. Observ.), while local fishermen try to play the game of 'mouse and cat' with the Park's Law Enforcement.

Catching versus Counting: Comparing the Pro-Environmental Attitudes, Behaviors, and Climate Concerns of Recreational Fishers and Citizen Scientists

The ongoing and interactive effects of climate change, overharvesting, and habitat loss on fish and fisheries impacts a wide array of stakeholders who rely on access to sustainable fish populations for their health, recreation, well-being, and income. Successful responses to these threats will require the involvement of stakeholders in co-developing solutions. Understanding the socio-psychological characteristics of these diverse stakeholders, including their environmental attitudes and behaviors, can potentially improve management support and effectiveness across and within these groups. Past research has focused on climate impacts and adaptation efforts in commercial fisheries, but less is known about climate-related perceptions and attitudes of recreational fishers and other stakeholders such as citizen scientists. This study investigated how pro-environmental attitudes (PEAs), pro-environmental behaviors (PEBs), and climate change concerns vary among fisheries-based recreationists based on activity type (recreational fishing, fish monitoring for citizen science) and specialization level. Among stakeholders, citizen scientists (fish counters) exhibited the strongest PEBs, followed by more specialized recreational fishers. Citizen scientists also had stronger PEAs than recreational fishers and non-fisher/non-citizen scientists, but there were no significant differences in PEAs across specialization levels. Citizen scientists showed greater concern for climate change than recreational fishers and non-fisher/non-citizen scientists. However, respondents overall showed greater concern for “the environment” than for climate change. Our results suggest that both activity type and specialization level are important considerations when developing strategies to promote pro-environmental behaviors and climate concerns and that communication frames centered on healthy ecosystems may be more effective with some stakeholders than those focused solely on climate change. We discuss implications for building public support of climate engagement efforts and sustainable, climate-resilient fisheries.

The seafood basket: Application of zero-inflated model to fish count purchase

The study analyzed demand for fish in the United States by the number of species purchased from a basket of species, rather than the usual approach focusing on one single species. The basket includes largemouth bass, hybrid striped bass, bluegill, walleye, yellow perch, rainbow trout, Great Lakes whitefish, lake trout, barramundi, and tilapia. These results largely confirm that consumers purchase different fish species or a combination of them from time to time and that consumers pay little attention to green factors such as sustainability, 3rd party certification and hormone-free when purchasing seafood. It appears that the effects of green credence attributes may not apply generically to a collection of fish or that the utility of green characteristics to seafood demand may be declining. As for production systems, although a higher percentage of consumers indicated preference for wild-caught fish, the importance of farmed fish to them was found to increase fish purchased.

LFCNet: A lightweight fish counting model based on density map regression

Accurate counting of fish is a fundamental task of fish farming and crucial for density regulation, bait planning, and economic benefit evaluation. However, computational capability is a substantial concern for practical aquaculture applications. The paper proposes a lightweight model especially for fish counting applications, termed LFCNet. The model is divided into three components: encoder, decoder, and generation head. Firstly, density map regression is utilized to address the high-density fish issue, making adhesion state counting more accurate. Secondly, Ghost modules are embedded to compress the parameters in mobile device applications. Finally, in order to realize the high-precision counting in a lightweight network, three concentrated-comprehensive convolution (C3) modules and transposed convolution layers are adopted to alleviate the computational overhead and recover the resolution of feature maps, respectively. Experiments show that parameters (16.26 MB) and floating point operations (FLOPs: 69.64G) of LFCNet are reduced by 73.8 % and 64.9 % compared with the baseline method. Meanwhile, the lower root mean square error (RMSE: 6.13) and mean absolute error (MAE: 4.10) indicate that LFCNet has higher counting precision and stability. The generalization studies demonstrate that LFCNet achieves the balance between accuracy and speed in various fish counting scenarios.

Microhabitat use of larval fish in a South Carolina Piedmont stream

Understanding habitat use and nursery areas of larval fish is a key component to managing and conserving riverine fishes. Yet, freshwater researchers often focus only on adult fishes, resulting in a limited understanding of the habitat requirements for the early life stages of freshwater fishes. The goal of this study was to quantify the larval fish microhabitat use of three fish families in Twelvemile Creek, a fifth-order tributary of Lake Hartwell (Savannah River basin) in the Piedmont ecoregion of South Carolina, USA. We used handheld dipnets to sample larval fishes along 20 equidistant transects spaced 10 m apart weekly from May to July 2021 along a 200 m stream reach. We also collected microhabitat data at each larval fish capture location. Most captured individuals were in the metalarval stage and were identified to the family level. A partial distance-based redundancy analysis indicated that water velocity contributed to changes in larval fish assemblage structure. Larval fishes occupied a subset of the available habitat that was characterized by low water velocity, non-Podostemum substrate, and shallow habitats close to the shore or bed rock structure. We also detected temporal patterns in larval fish counts, with peak Percidae and Leuciscidae counts in late July and the highest Catostomidae counts in late May–early June. Our results suggest that larval fishes select habitats with low water velocity and shallow habitats close to shore microhabitat characteristics, and that riffle-pool sequences may serve as a nursery habitat for Percidae, Catostomidae and Leuciscidae metalarvae.

Robust detection of farmed fish by fusing YOLOv5 with DCM and ATM

Accurate object detection of farmed fish is vital for fish counting and measuring. However, the blurring, occlusion, and varying density of underwater fish images from aquaculture ponds result in low precision and poor robustness. This paper proposes a robust detection model called DCM-ATM-YOLOv5, in which the deformable convolution module (DCM) and adaptive threshold module (ATM) are integrated into YOLOv5. The farmed fish detection model is based on YOLOv5, considering its high precision and real-time capabilities. To achieve higher precision on blurred targets, the DCM is applied to offset the sampling locations and to focus more on fish so that the features of blurry fish are enhanced and the background is suppressed. The ATM is introduced to alleviate missed detections resulting from fixed thresholds by generating appropriate thresholds for different densities and predicting the thresholds dynamically, improving the robustness of the proposed model. Two experiments based on data from aquaculture ponds were designed to test the performance of the proposed model. The average precision and recall of DCM-ATM-YOLOv5 were 97.53% and 98.09%, respectively. Compared with the average precision and recall of YOLOv5, those of DCM-ATM-YOLOv5 were improved by 1.92% and 1.97%, respectively. The robustness of the proposed model was evaluated based on data with different densities and settings. The detection results obtained from datasets of varying densities and settings demonstrated that the detection performance of DCM-ATM-YOLOv5 was still superior to that of YOLOv5, indicating that the proposed model has significant stability and robustness. The experimental results showed that DCM-ATM-YOLOv5 has high precision and strong robustness in genuine aquaculture settings and can provide a feasible solution for the detection of farmed fish.

Effect of Anisakis pegreffii (Nematoda: Anisakidae) on Biochemical and Haematological Characteristics of Chub Mackerel (Scomber japonicus Houttuyn, 1782) Caught in the Dardanelles at Çanakkale, Türkiye

Chub mackerel (Scomber japonicus Houttuyn, 1782) were obtained from the commercial fishermen in Çanakkale, Turkey, in July, 2017. We examined a total of 40 fish (20 non-infested and 20 infested) and assessed the biometric indices, haematological parameters and serum biochemical variables. The hepatosomatic index and gonadosomatic index of infested chub mackerel fish were lower than those of non-infested fish. Blood haematocrit ratio and haemoglobin concentration in naturally parasite-infested chub mackerel fish were significantly lower than those in non-infested fish. However, white blood cell counts of the parasite-infested chub mackerel fish were higher than those of healthy ones. Serum total protein, globulin, glucose, cholesterol, triglyceride, urea, chlorine and iron levels in naturally parasite-infested chub mackerel fish were significantly lower than those in non-infested fish. Moreover, serum lactate dehydrogenase, aspartate aminotransferase, alkaline phosphatase and alanine aminotransferase activities of the parasite-infested chub mackerel fish were higher than those in healthy ones. Therefore, observed variations in haematological parameters, serum biochemical variables and biometric indices influenced by the parasite, A. pegreffii, may potentially increase sensitivity of the chub mackerel, Scomber japonicus, to diseases and environmental conditions.

Environmentally adaptive fish or no-fish classification for river video fish counters using high-performance desktop and embedded hardware

Automated fish counters featuring robust, real-time computer vision capabilities can provide a cost-effective means to count migrating freshwater fish. In this work, we propose a four-stage process for automatically sorting videos with and without fish. Underwater fish counter videos provide a challenging range of environmental conditions including clear water, biofilm growth, bubbles, turbidity, low light and overexposure. To address this, our method also includes the automated classification of these six environmental conditions. The proposed methods are computationally efficient and can be implemented on servers, high-performance desktop computers and low-cost, energy-efficient embedded hardware. The models were trained, tested, and validated using a collection of 3000 videos taken from underwater fish counter installations in several alpine and lowland European rivers provided by commercial and governmental collaborators. This work demonstrates a fast, accurate, and robust computer vision workflow for large-scale automated freshwater fish counting systems.

Characterization and Isolation of Bacteria from Adult African Catfish (Clarias gariepinus) in Selected Fish Farms of Rivers State, Nigeria

This research examined the bacterial loads in the intestine, skin and gills of cultured Clarias gariepinus in some selected private farms at Eleme, Emohua, Ikwerre, Obio/Akpor and Port Harcourt Local Government Areas of Rivers State, Nigeria. Determination of bacterial loads species characterization in fish parts was done using standard microbiological procedures. Results of the assay showed high bacterial count in the gill, gut and skin of fish from the different locations, especially at Emuoha Local Government Area. The total heterotrophic bacteria counts of fish: gills ranged between 1.7X108 and 3.1 X 107; gut ranged between 1.4 X 107 and 7.2 X 106 and skin ranged between 1.5 X 106 and 6.8 X 106 in all the stations under consideration (Eleme, Emohua, Ikwerre, Obio/Akpor and Port Harcourt city Local Government Areas of Rivers State, Nigeria.). Bacterial genera isolated includs: Escherichia coli, Salmonella sp. Enterica sp., Staphylococcus sp., Bacillus sp., Shigella sp., Enterobacter sp., Klebsiella sp. and Citrobacter sp.

Machine learning applied to big data from marine cabled observatories: A case study of sablefish monitoring in the NE Pacific

Ocean observatories collect large volumes of video data, with some data archives now spanning well over a few decades, and bringing the challenges of analytical capacity beyond conventional processing tools. The analysis of such vast and complex datasets can only be achieved with appropriate machine learning and Artificial Intelligence (AI) tools. The implementation of AI monitoring programs for animal tracking and classification becomes necessary in the particular case of deep-sea cabled observatories, as those operated by Ocean Networks Canada (ONC), where Petabytes of data are now collected each and every year since their installation. Here, we present a machine-learning and computer vision automated pipeline to detect and count sablefish (Anoplopoma fimbria), a key commercially exploited species in the N-NE Pacific. We used 651 hours of video footage obtained from three long-term monitoring sites in the NEPTUNE cabled observatory, in Barkley Canyon, on the nearby slope, and at depths ranging from 420 to 985 m. Our proposed AI sablefish detection and classification pipeline was tested and validated for an initial 4.5 month period (Sep 18 2019-Jan 2 2020), and was a first step towards validation for future processing of the now decade-long video archives from Barkley Canyon. For the validation period, we trained a YOLO neural network on 2917 manually annotated frames containing sablefish images to obtain an automatic detector with a 92% Average Precision (AP) on 730 test images, and a 5-fold cross-validation AP of 93% (± 3.7%). We then ran the detector on all video material (i.e., 651 hours from a 4.5 month period), to automatically detect and annotate sablefish. We finally applied a tracking algorithm on detection results, to approximate counts of individual fishes moving on scene and obtain a time series of proxy sablefish abundance. Those proxy abundance estimates are among the first to be made using such a large volume of video data from deep-sea settings. We discuss our AI results for application on a decade-long video monitoring program, and particularly with potential for complementing fisheries management practices of a commercially important species.

Computer vision in aquaculture: a case study of juvenile fish counting

ABSTRACT In aquaculture breeding or production programmes, counting juvenile fish represents a considerable cost in terms of the human hours needed. In this study, we explored the use of two state-of-the-art machine learning architectures (Single Shot Detection, hereafter SSD and Faster Regions with convolutional neural networks, hereafter Faster R-CNN) to augment a manual image-based juvenile fish counting method for the Australasian snapper (Chrysophrys auratus) bred at The New Zealand Institute for Plant and Food Research Limited. We tested model accuracy after tuning for confidence thresholds and non-maximal suppression overlap parameters, and implementing a bias correction using a Poisson regression model. Validation of image data showed that after tuning, bias-corrected SSD and Faster R-CNN models had mean absolute percent errors (MAPE) of less than 10%, with SSD having MAPE of less than 5%. Comparison of the results with those from manual counts showed that, while manual counts are slightly more accurate (MAPE = 1.56), the machine learning methods allow for more rapid assessment of counts and thus facilitating a higher throughput. This work represents a first step for deploying machine learning applications to an existing real-life aquaculture scenario and provides a useful starting point for further developments, such as real-time counting of fish or collecting additional phenotypic data from the source images.

The role of nocturnal fishes on coral reefs: A quantitative functional evaluation.

The ecological functions of nocturnal coral reef fishes are poorly known. Yet, nocturnal resources for coral reef consumers are theoretically as abundant and productive, if not more so, than their diurnal counterparts. In this study, we quantify and contrast the energetic dynamics of nocturnal and diurnal fishes in a model coral reef ecosystem, evaluating whether they attain similar levels of biomass production. We integrated a detailed dataset of coral reef fish counts, comprising diurnal and nocturnal species, in sites sheltered and exposed to wave action. We combined somatic growth and mortality models to estimate rates of consumer biomass production, a key ecosystem function. We found that diurnal fish assemblages have a higher biomass than nocturnal fishes: 104% more in sheltered sites and 271% more in exposed sites. Differences in productivity were even more pronounced, with diurnal fishes contributing 163% more productivity in sheltered locations, and 558% more in exposed locations. Apogonidae dominated biomass production within the nocturnal fish assemblage, comprising 54% of total nocturnal fish productivity, which is proportionally more than any diurnal fish family. The substantially lower contributions of nocturnal fishes to biomass and biomass production likely indicate constraints on resource accessibility. Taxa that overcome these constraints may thrive, as evidenced by apogonids. This study highlights the importance of nocturnal fishes in underpinning the flow of energy and nutrients from nocturnal resources to reef communities; a process driven mainly by small, cryptic fishes.

Sampling errors and variability in video transects for assessment of reef fish assemblage structure and diversity.

Video monitoring is a rapidly evolving tool in aquatic ecological research because of its non-destructive ability to assess fish assemblages. Nevertheless, methodological considerations of video monitoring techniques are often overlooked, especially in more complex sampling designs, causing inefficient data collection, processing, and interpretation. In this study, we discuss how video transect sampling designs could be assessed and how the inter-observer variability, design errors and sampling variability should be quantified and accounted for. The study took place in the coastal areas of the Galapagos archipelago and consisted of a hierarchical repeated-observations sampling design with multiple observers. Although observer bias was negligible for the assessment of fish assemblage structure, diversity and counts of individual species, sampling variability caused by simple counting/detection errors, observer effects and instantaneous fish displacement was often important. Especially for the counts of individual species, sampling variability most often exceeded the variability of the transects and sites. An extensive part of the variability in the fish assemblage structure was explained by the different transects (13%), suggesting that a sufficiently high number of transects is required to account for the within-location variability. Longer transect lengths allowed a better representation of the fish assemblages as sampling variability decreased by 33% if transect length was increased from 10 to 50 meters. However, to increase precision, including more repeats was typically more efficient than using longer transect lengths. The results confirm the suitability of the technique to study reef fish assemblages, but also highlight the importance of a sound methodological assessment since different biological responses and sampling designs are associated with different levels of sampling variability, precision and ecological relevance. Therefore, besides the direct usefulness of the results, the procedures to establish them may be just as valuable for researchers aiming to optimize their own sampling technique and design.

Estimating length composition of fish observed with stereo-video cameras: A simulation study with application to red snapper (Lutjanus campechanus)

Stereo-video cameras have become an important tool worldwide for enumerating abundance and length compositions of marine fish. The two most common approaches for enumerating are referred to as MaxN and MeanCount, where the former counts fish on the single video frame with the most individuals observed and the latter uses the mean across multiple frames. Previous studies have demonstrated that both approaches may work well for computing relative abundance. However, basic fish population age-structure (e.g. more younger fish) and fish schooling characteristics (e.g. fish of similar size/age swim together) suggest the potential for bias in length compositions computed from the MaxN approach, as the single video frame with the most individuals would tend to overrepresent smaller, more populous fish. To evaluate the two approaches, we simulated a stationary video system placed on a sampling site inhabited by moving fish, and we compared length compositions estimated from each approach to that of the true, underlying population. Indeed, MaxN appears biased over a range of potential factors, in which the bias is toward estimates of mean size that are smaller than the true values. The factors leading to biased age/length estimates from MaxN include 1) large fish counts in the single frame, 2) fish schooling by age/length, especially at moderate school densities, and 3) small camera viewing area relative to the site being sampled. In addition to the general evaluation, we applied the simulation framework to video parameters derived from data on red snapper (Lutjanus campechanus) in U.S. Atlantic waters. The results suggest that the MaxN length composition bias is small for red snapper due primarily to the low counts of fish on site and low levels of schooling observed.

Method of quantitative visual assessment of abundance of sockeye salmon spawners running to Kurile Lake (Kamchatka) for reproduction

Original thechnique of counting Pacific salmon on fish counting facility located at the head of the Ozernaya River, at KamchatNIRO survey station, is presented for the first time.

Distinct microplastic patterns in the sediment and biota of an urban stream

Urban freshwaters, their sediments and resident biota are often highly susceptible to microplastic contamination from catchment-specific sources. Water velocity and spatiotemporal dynamics within the system can impact microplastic loads, while biological features may additionally impact levels within freshwater biota. Here, we investigated the spatiotemporal variations in microplastic loads collected from sediment, macroinvertebrate and fish samples from an urban watercourse (Bourne Stream) in Dorset, southwest England. Sediment particles were mostly fragments of colours (especially orange and purple) whereas microplastics in both macroinvertebrates and fishes were blue/green and fibres. Across all sample types, the dominant particle size class was ≤100 μm. Median (M) and range (R) of microplastic loads within each sample type were sediment: M = 0.06, R = 0–0.36 particles g−1; macroinvertebrates: M = 0, R = 0–4 particles per batch; and fishes: M = 1, R = 0–6 particles per individual. Sediment loads varied spatially, with the highest load in the most upstream site, whereas biotic loads did not vary across space and time. Macroinvertebrate batch loadings varied between taxa and feeding guild, with counts significantly higher in annelids but lower in herbivores. Fish counts were higher in species with true, differentiated stomachs, but with the effects of species, feeding guild and body size being non-significant. Within sites, mean microplastic loads did not correlate between sediment, macroinvertebrate and fish samples. These results suggest that sediment freshwater microplastic loadings may vary spatially but that these trends are not reflected by, or correlated to, those in the biota where ingestion varies with biological traits. Assessments of freshwater microplastic contamination must therefore consider sampling spatiotemporally and across different biotic communities to fully understand the scale of contamination, and to subsequently undertake effective mitigation steps.

Nondestructive determination of the freshness change in bighead carp heads under variable temperatures by using excitation-emission matrix fluorescence and back-propagation neural networks

This study established back-propagation neural networks (BPNNs) for evaluating the freshness of bighead carp ( Hypophthalmichthys nobilis ) heads during chilled storage via fluorescence spectroscopy using an excitation-emission matrix (EEM). The total volatile basic nitrogen (TVB-N) and total aerobic count (TAC) of fish increased obviously during storage at 0, 4, 8, 12, and 16 °C, while sensory scores decreased with increasing storage time. The EEM fluorescence intensity was measured, and its change was correlated with the freshness indicators of the samples. Three characteristic components of EEM data were extracted by parallel factor analysis, and two freshness indicators were used to construct the EEM-BPNNs model. The results demonstrated that the relative errors of the EEM-BPNNs model for TVB-N and TAC were less than 14%. This result indicated that the EEM-BPNNs model could determine the freshness of fish in cold chains in a rapid and nondestructive way.

Microplastic loads within riverine fishes and macroinvertebrates are not predictable from ecological or morphological characteristics

Microplastics are a relatively new but important form of freshwater contamination that can be ingested by a range of different species, with particle counts thought to be predictable from species ecology and morphology. Here, we report levels of microplastics in a 26 μm-5 mm size range within the macroinvertebrate and fish community of a lowland river (Dorset Stour, SW England), and test the hypothesis that counts are predictable from characteristics such as feeding guild, body length and trophic position. Macroinvertebrates (n = 257, 12 taxa) and fish (n = 418, 9 species) were collected from distinct river reaches by kick sampling and rod and line angling, respectively. Batches of whole macroinvertebrates and individual fish gastrointestinal tracts were digested with 30% hydrogen peroxide before microplastic screening and FTIR polymer confirmation on a particle subset. Particles were found in 40% of pooled macroinvertebrate batches (taxa incidences: 14–75%) and 39% of fishes (species incidences: 29–47%). Dominant particle feature categories were ≤100 μm, blue/green, fragments and fibres identified as various polyolefins. Although particle counts in macroinvertebrates were highest in Ephemeroptera (mean of 0.74 particles per individual), the relationships between particle loads, batch number and guild were all non-significant. In fishes, particle counts were not significantly related to species, stomach structure, feeding guild or body length, with spatial differences also not apparent across the catchment. Individual fish particle counts were similarly not significantly associated with their trophic positions (calculated from bulk δ15N values for a subset of fishes) and parasite load of Pomphorhynchus tereticollis. Correlations between fish and macroinvertebrate particle counts within specific river reaches were also not significant. In entirety, these results indicated although loadings of microplastic particles were relatively consistent within the two communities, they were not predictable from any of their ecological or morphological characteristics.

Long-Term Monitoring of Diel and Seasonal Rhythm of Dentex dentex at an Artificial Reef

Behavioral rhythms are a key aspect of species fitness, since optimize ecological activities of animals in response to a constantly changing environment. Cabled observatories enable researchers to collect long-term biological and environmental data in real-time, providing relevant information on coastal fishes’ ecological niches and their temporal regulation (i.e., phenology). In this framework, the platform OBSEA (an EMSO Testing-Site in the NW coastal Mediterranean) was used to monitor the 24-h and seasonal occurrence of an ecologically iconic (i.e., top-predator) coastal fish species, the common dentex (Dentex dentex). By coupling image acquisition with oceanographic and meteorological data collection at a high-frequency (30 min), we compiled 8-years’ time-series of fish counts, showing daytime peaks by waveform analysis. Peaks of occurrence followed the photophase limits as an indication of photoperiodic regulation of behavior. At the same time, we evidenced a seasonal trend of counts variations under the form of significant major and minor increases in August and May, respectively. A progressive multiannual trend of counts increase was also evidenced in agreement with the NW Mediterranean expansion of the species. In GLM and GAM modeling, counts not only showed significant correlation with solar irradiance but also with water temperature and wind speed, providing hints on the species reaction to projected climate change scenarios. Grouping behavior was reported mostly at daytime. Results were discussed assuming a possible link between count patterns and behavioral activity, which may influence video observations at different temporal scales.

Automatic fish counting via a multi-scale dense residual network

Automatic fish counting via a multi-scale dense residual network