Application of machine learning algorithms to identify cryptic reproductive habitats using diverse information sources

Fishery catch records support machine learning-based prediction of illegal fishing off US West Coast

Illegal, unreported, and unregulated (IUU) fishing is a major problem worldwide, often made more challenging by a lack of at-sea and shoreside monitoring of commercial fishery catches. Off the US West Coast, as in many places, a primary concern for enforcement and management is whether vessels are illegally fishing in locations where they are not permitted to fish. We explored the use of supervised machine learning analysis in a partially observed fishery to identify potentially illicit behaviors when vessels did not have observers on board. We built classification models (random forest and gradient boosting ensemble tree estimators) using labeled data from nearly 10,000 fishing trips for which we had landing records (i.e., catch data) and observer data. We identified a set of variables related to catch (e.g., catch weights and species) and delivery port that could predict, with 97% accuracy, whether vessels fished in state versus federal waters. Notably, our model performances were robust to inter-annual variability in the fishery environments during recent anomalously warm years. We applied these models to nearly 60,000 unobserved landing records and identified more than 500 instances in which vessels may have illegally fished in federal waters. This project was developed at the request of fisheries enforcement investigators, and now an automated system analyzes all new unobserved landings records to identify those in need of additional investigation for potential violations. Similar approaches informed by the spatial preferences of species landed may support monitoring and enforcement efforts in any number of partially observed, or even totally unobserved, fisheries globally.

Telemetry-based spatial-temporal fish habitat models for fishes in an urban freshwater harbour

Fish habitat associations are important measures for effective aquatic habitat management, but often vary over broad spatial and temporal scales, and are therefore challenging to measure comprehensively. We used a 9-year acoustic telemetry dataset to generate spatial-temporal habitat suitability models for seven fish species in an urban freshwater harbour, Toronto Harbour, Lake Ontario. Fishes generally occupied the more natural regions of Toronto Harbour most frequently. However, each species exhibited unique habitat associations and spatial-temporal interactions in their habitat use. For example, largemouth bass exhibited the most consistent seasonal habitat use, mainly associating with shallow, sheltered embayments with high aquatic vegetation (SAV) cover. Conversely, walleye seldom occupied Toronto Harbour in summer, with the highest occupancy of shallow, low-SAV habitats in the spring, which corresponds to their spawning period. Others, such as common carp, shifted between shallow summer and deeper winter habitats. Community level spatial-temporal habitat importance estimates were also generated, which can serve as an aggregate measure for habitat management. Acoustic telemetry provides novel opportunities to generate robust spatial-temporal fish habitat models based on wild fish behaviour, which are useful for the management of fish habitat from a fish species and community perspective.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s10750-023-05180-z.

The movement ecology of fishes

Movement of fishes in the aquatic realm is fundamental to their ecology and survival. Movement can be driven by a variety of biological, physiological and environmental factors occurring across all spatial and temporal scales. The intrinsic capacity of movement to impact fish individually (e.g., foraging) with potential knock-on effects throughout the ecosystem (e.g., food web dynamics) has garnered considerable interest in the field of movement ecology. The advancement of technology in recent decades, in combination with ever-growing threats to freshwater and marine systems, has further spurred empirical research and theoretical considerations. Given the rapid expansion within the field of movement ecology and its significant role in informing management and conservation efforts, a contemporary and multidisciplinary review about the various components influencing movement is outstanding. Using an established conceptual framework for movement ecology as a guide (i.e., Nathan et al., 2008: 19052), we synthesized the environmental and individual factors that affect the movement of fishes. Specifically, internal (e.g., energy acquisition, endocrinology, and homeostasis) and external (biotic and abiotic) environmental elements are discussed, as well as the different processes that influence individual-level (or population) decisions, such as navigation cues, motion capacity, propagation characteristics and group behaviours. In addition to environmental drivers and individual movement factors, we also explored how associated strategies help survival by optimizing physiological and other biological states. Next, we identified how movement ecology is increasingly being incorporated into management and conservation by highlighting the inherent benefits that spatio-temporal fish behaviour imbues into policy, regulatory, and remediation planning. Finally, we considered the future of movement ecology by evaluating ongoing technological innovations and both the challenges and opportunities that these advancements create for scientists and managers. As aquatic ecosystems continue to face alarming climate (and other human-driven) issues that impact animal movements, the comprehensive and multidisciplinary assessment of movement ecology will be instrumental in developing plans to guide research and promote sustainability measures for aquatic resources.

Predator-prey landscapes of large sharks and game fishes in the Florida Keys

Interspecific interactions can play an essential role in shaping wildlife populations and communities. To date, assessments of interspecific interactions, and more specifically predator-prey dynamics, in aquatic systems over broad spatial and temporal scales (i.e., hundreds of kilometers and multiple years) are rare due to constraints on our abilities to measure effectively at those scales. We applied new methods to identify space-use overlap and potential predation risk to Atlantic tarpon (Megalops atlanticus) and permit (Trachinotus falcatus) from two known predators, great hammerhead (Sphyrna mokarran) and bull (Carcharhinus leucas) sharks, over a 3-year period using acoustic telemetry in the coastal region of the Florida Keys (USA). By examining spatiotemporal overlap, as well as the timing and order of arrival at specific locations compared to random chance, we show that potential predation risk from great hammerhead and bull sharks to Atlantic tarpon and permit are heterogeneous across the Florida Keys. Additionally, we find that predator encounter rates with these game fishes are elevated at specific locations and times, including a prespawning aggregation site in the case of Atlantic tarpon. Further, using machine learning algorithms, we identify environmental variability in overlap between predators and their potential prey, including location, habitat, time of year, lunar cycle, depth, and water temperature. These predator-prey landscapes provide insights into fundamental ecosystem function and biological conservation, especially in the context of emerging fishery-related depredation issues in coastal marine ecosystems.

Application of machine learning and acoustic predation tags to classify migration fate of Atlantic salmon smolts

Mortality and predation of tagged fishes present a serious challenge to interpreting results of acoustic telemetry studies. There is a need for standardized methods to identify predated individuals and reduce the impacts of "predation bias" on results and conclusions. Here, we use emerging approaches in machine learning and acoustic tag technology to classify out-migrating Atlantic salmon (Salmo salar) smolts into different fate categories. We compared three methods of fate classification: predation tag pH sensors and detection data, unsupervised k-means clustering, and supervised random forest combined with tag pH sensor data. Random forest models increased predation estimates by 9-32% compared to relying solely on pH sensor data, while clustering reduced estimates by 3.5-30%. The greatest changes in fate class estimates were seen in years with large class imbalance (one or more fate classes underrepresented compared to the others) or low model accuracy. Both supervised and unsupervised approaches were able to classify smolt fate; however, in-sample model accuracy improved when using tag sensor data to train models, emphasizing the value of incorporating such sensors when studying small fish. Sensor data may not be sufficient to identify predation in isolation due to Type I and Type II error in predation sensor triggering. Combining sensor data with machine learning approaches should be standard practice to more accurately classify fate of tagged fish.

Unsupervised Machine Learning and Data Mining Procedures Reveal Short Term, Climate Driven Patterns Linking Physico-Chemical Features and Zooplankton Diversity in Small Ponds

Machine Learning (ML) is an increasingly accessible discipline in computer science that develops dynamic algorithms capable of data-driven decisions and whose use in ecology is growing. Fuzzy sets are suitable descriptors of ecological communities as compared to other standard algorithms and allow the description of decisions that include elements of uncertainty and vagueness. However, fuzzy sets are scarcely applied in ecology. In this work, an unsupervised machine learning algorithm, fuzzy c-means and association rules mining were applied to assess the factors influencing the assemblage composition and distribution patterns of 12 zooplankton taxa in 24 shallow ponds in northern Italy. The fuzzy c-means algorithm was implemented to classify the ponds in terms of taxa they support, and to identify the influence of chemical and physical environmental features on the assemblage patterns. Data retrieved during 2014 and 2015 were compared, taking into account that 2014 late spring and summer air temperatures were much lower than historical records, whereas 2015 mean monthly air temperatures were much warmer than historical averages. In both years, fuzzy c-means show a strong clustering of ponds in two groups, contrasting sites characterized by different physico-chemical and biological features. Climatic anomalies, affecting the temperature regime, together with the main water supply to shallow ponds (e.g., surface runoff vs. groundwater) represent disturbance factors producing large interannual differences in the chemistry, biology and short-term dynamic of small aquatic ecosystems. Unsupervised machine learning algorithms and fuzzy sets may help in catching such apparently erratic differences.

Ciguatera Mini Review: 21st Century Environmental Challenges and the Interdisciplinary Research Efforts Rising to Meet Them

Globally, the livelihoods of over a billion people are affected by changes to marine ecosystems, both structurally and systematically. Resources and ecosystem services, provided by the marine environment, contribute nutrition, income, and health benefits for communities. One threat to these securities is ciguatera poisoning; worldwide, the most commonly reported non-bacterial seafood-related illness. Ciguatera is caused by the consumption of (primarily) finfish contaminated with ciguatoxins, potent neurotoxins produced by benthic single-cell microalgae. When consumed, ciguatoxins are biotransformed and can bioaccumulate throughout the food-web via complex pathways. Ciguatera-derived food insecurity is particularly extreme for small island-nations, where fear of intoxication can lead to fishing restrictions by region, species, or size. Exacerbating these complexities are anthropogenic or natural changes occurring in global marine habitats, e.g., climate change, greenhouse-gas induced physical oceanic changes, overfishing, invasive species, and even the international seafood trade. Here we provide an overview of the challenges and opportunities of the 21st century regarding the many facets of ciguatera, including the complex nature of this illness, the biological/environmental factors affecting the causative organisms, their toxins, vectors, detection methods, human-health oriented responses, and ultimately an outlook towards the future. Ciguatera research efforts face many social and environmental challenges this century. However, several future-oriented goals are within reach, including digital solutions for seafood supply chains, identifying novel compounds and methods with the potential for advanced diagnostics, treatments, and prediction capabilities. The advances described herein provide confidence that the tools are now available to answer many of the remaining questions surrounding ciguatera and therefore protection measures can become more accurate and routine.