We propose a fish detection system based on deep network architectures to robustly detect and count fish objects under a variety of benthic background and illumination conditions. The algorithm consists of an ensemble of Region-based Convolutional Neural Networks that are linked in a cascade structure by Long Short-Term Memory networks. The proposed network is efficiently trained as all components are jointly trained by backpropagation. We train and test our system for a dataset of 18 videos taken in the wild. In our dataset, there are around 20 to 100 fish objects per frame with many fish objects having small pixel areas (less than 900 square pixels). From a series of experiments and ablation tests, the proposed system preserves detection accuracy despite multi-scale distortions, cropping and varying background environments. We present analysis that shows how object localization accuracy is increased by an automatic correction mechanism in the deep network's cascaded ensemble structure. The correction mechanism rectifies any errors in the predictions as information progresses through the network cascade. Our findings in this experiment regarding ensemble system architectures can be generalized to other object detection applications.
Tools and Data
The Bay of Biscay is being affected by increasing level of marine litter, which is causing a wide variety of adverse environmental, social, public health, safety and economic impacts. The term “beach littering” has been coined to refer to the marine litter that is deposited on beaches. This litter may come from the sea and through land-based pathways, either from remote or adjacent areas. Dirty beaches can derive in loss of aesthetical value, beach cleaning cost, environmental harm or tourism revenue reduction among others. Therefore, local authorities have started to search for cost-effective approaches to understand and reduce litter accumulation in their beaches. A model is presented in this paper, which is based on Bayesian Networks and enables the forecasting of marine litter beaching at seven beaches located on the south-eastern coast of the Bay of Biscay. The model uses 9.5 years of metocean, environmental and beach cleaning data. The class to predict was defined as a variable with two possible values: Low and High accumulation of beach litter. The obtained models reached an average accuracy of 65.3 ± 6.4%, being the river flow, precipitation, wind and wave the most significant predictors and likely drivers of litter accumulation in beaches. These models may provide some insight to local authorities on the drivers affecting the litter beaching and may help to define their strategies for its reduction.
Issues related to protection of the Arctic environment are becoming increasingly urgent, as arctic ecosystems are vulnerable to increasing anthropogenic pressures. The problem of protecting Northern nature from the effects of persistent organic pollutants, which are dangerous for both biota and human health, is particularly acute. This case study analyses the existing normative acts regulating monitoring activities in the Russian Arctic. The paper emphasizes gaps in legal regulation, which are particularly prominent with regard to monitoring the quality of traditional food consumed by indigenous peoples. The author introduces proposals to change the current legislation to improve the efficiency of the state monitoring system in the Russian Arctic. Such changes will also help to harmonize monitoring activities in Russia with other Arctic States and to fill in the gaps in the Global Monitoring Reports and the Arctic Monitoring and Assessment Programme (AMAP) reports on persistent organic pollutants in traditional indigenous food.
Digital photography is widely used by coral reef monitoring programs to assess benthic status and trends. In addition to creating a permanent archive, photographic surveys can be rapidly conducted, which is important in environments where bottom-time is frequently limiting. However, substantial effort is required to manually analyze benthic images; which is expensive and leads to lags before data are available. Using previously analyzed imagery from NOAA’s Pacific Reef Assessment and Monitoring Program, we assessed the capacity of a trained and widely used machine-learning image analysis tool – CoralNet coralnet.ucsd.edu – to generate fully-automated benthic cover estimates for the main Hawaiian Islands (MHI) and American Samoa. CoralNet was able to generate estimates of site-level coral cover for both regions that were highly comparable to those generated by human analysts (Pearson’s r > 0.97, and with bias of 1% or less). CoralNet was generally effective at estimating cover of common coral genera (Pearson’s r > 0.92 and with bias of 2% or less in 6 of 7 cases), but performance was mixed for other groups including algal categories, although generally better for American Samoa than MHI. CoralNet performance was improved by simplifying the classification scheme from genus to functional group and by training within habitat types, i.e., separately for coral-rich, pavement, boulder, or “other” habitats. The close match between human-generated and CoralNet-generated estimates of coral cover pooled to the scale of island and year demonstrates that CoralNet is capable of generating data suitable for assessing spatial and temporal patterns. The imagery we used was gathered from sites randomly located in <30 m hard-bottom at multiple islands and habitat-types per region, suggesting our results are likely to be widely applicable. As image acquisition is relatively straightforward, the capacity of fully-automated image analysis tools to minimize the need for resource intensive human analysts opens possibilities for enormous increases in the quantity and consistency of coral reef benthic data that could become available to researchers and managers.
Instrumented moorings (hereafter referred to as moorings), which are anchored buoys or an anchored configuration of instruments suspended in the water column, are highly valued for their ability to host a variety of interchangeable oceanographic and meteorological sensors. This flexibility makes them a useful technology for meeting end user and science-driven requirements. Overall, societal needs related to human health, safety, national security, and economic prosperity in coastal areas are met through the availability of continuous data from coastal moorings and other complementary observing platforms within the Earth-observing system. These data streams strengthen the quality and accuracy of data products that inform the marine transportation industry, the tourism industry, fisheries, the military, public health officials, coastal and emergency managers, educators, and research scientists, among many others. Therefore, it is critical to sustain existing observing system networks, especially during this time of extreme environmental variability and change. Existing fiscal and operational challenges affecting the sustainability of observing networks will likely continue into the next decade, threatening the quality of downstream data and information products – especially those used for long-term monitoring, planning, and decision-making. This paper describes the utility of coastal moorings as part of an integrated coastal observing system, with an emphasis on stakeholder engagement to inform observing requirements and to ensure data products are tailored to user needs. We provide 10 recommendations for optimizing moorings networks, and thus downstream data products, to guide regional planners, and network operators:
1. Develop strategies to increase investment in coastal mooring networks
2. Collect stakeholder priorities through targeted and continuous stakeholder engagements
3. Include complementary systems and emerging technologies in implementation planning activities
4. Expand and sustain water column ecosystem moorings in coastal locations
5. Coordinate with operators and data managers across geographic scales
6. Standardize and integrate data management best practices
7. Provide open access to data
8. Promote environmental health and operational safety stewardship and regulatory compliance
9. Develop coastal mooring observing network performance metrics
10. Routinely monitor and assess the design of coastal mooring networks
Sea state information is needed for many applications, ranging from safety at sea and on the coast, for which real time data are essential, to planning and design needs for infrastructure that require long time series. The definition of the wave climate and its possible evolution requires high resolution data, and knowledge on possible drift in the observing system. Sea state is also an important climate variable that enters in air-sea fluxes parameterizations. Finally, sea state patterns can reveal the intensity of storms and associated climate patterns at large scales, and the intensity of currents at small scales. A synthesis of user requirements leads to requests for spatial resolution at kilometer scales, and estimations of trends of a few centimeters per decade. Such requirements cannot be met by observations alone in the foreseeable future, and numerical wave models can be combined with in situ and remote sensing data to achieve the required resolution. As today's models are far from perfect, observations are critical in providing forcing data, namely winds, currents and ice, and validation data, in particular for frequency and direction information, and extreme wave heights. In situ and satellite observations are particularly critical for the correction and calibration of significant wave heights to ensure the stability of model time series. A number of developments are underway for extending the capabilities of satellites and in situ observing systems. These include the generalization of directional measurements, an easier exchange of moored buoy data, the measurement of waves on drifting buoys, the evolution of satellite altimeter technology, and the measurement of directional wave spectra from satellite radar instruments. For each of these observing systems, the stability of the data is a very important issue. The combination of the different data sources, including numerical models, can help better fulfill the needs of users.
In recent years, coral reef degradation has been increasing. Management and conservation efforts have tended to focus only on the physical condition of the coral reefs with less attention to biological and oceanographic aspects, in particular genetics and hydrodynamics. Genetic data can illustrate the connectivity between and within populations of an organism, making it is possible to determine source and sink populations or sites. Studies of physical water movements can also illustrate the likely patterns of movement or predict the mobility of coral planulae. Both of these approaches can help to strengthen Marine Protected Area (MPA) design, especially at the formation stage, in particular MPAs focused on coral reef ecosystems. Together, these two approaches can provide data on biological networks in a region and help delineate stocks. The implications of such studies can help to identify conservation priorities and improve the effectiveness of management processes in Indonesia, and can certainly enable the refinement of general approaches to help produce management plans tailored to local and regional conditions and processes. This brief review aims to review the constraints that occur in the management process, including barriers to and potential benefits of integrating molecular and hydrodynamic data into the management and conservation process, as illustrated through a critical review of MPA implementation in the waters around Sulawesi Island.
Applying a proteomic approach for biomonitoring marine environments offers a useful tool for identifying organisms’ stress responses, with benthic filter-feeders being ideal candidates for this practice. Here, we investigated the proteomic profile of two solitary ascidians (Chordata, Ascidiacea): Microcosmus exasperatus, collected from five sites along the Mediterranean coast of Israel; and Polycarpa mytiligera collected from four sites along the Red Sea coast. 193 and 13 proteins in M. exasperatus and P. mytiligera, respectively, demonstrated a significant differential expression. Significant differences were found between the proteomes from the northern and the southern sites along both the Mediterranean and the Red Sea coasts. Some of the significant proteins had previously been shown to be affected by environmental stressors, and thus have the potential to be further developed as biomarkers. Obtaining a proteomic profile of field-collected ascidians provides a useful tool for the early-detection of a stress response in ascidians worldwide.
The currently existing Biological Protection Areas (BPAs) are Italian conservation measures specifically oriented to preserve/recover commercial fish stocks through the regulation or ban of few fishing activities. Although BPAs have been well identified/designed within the Italian national waters, neither data on their effects on fishery resources nor on the fishers’ compliance with these no-take regulations are yet available.
In this context, the present study was aimed to investigate how AIS data processing could be used to map patterns in fishing activity within/around small regulated areas and rate the effectiveness of the conservation measures in place. To do this, it was impossible not to address the issue surrounding well-known and substantial gaps in data coverage and attempt an estimate of them by flagging events when AIS had been switched off within/around BPAs.
Results highlighted that almost all the BPAs are illegally trawled and that, unless additional legislation is effected to regulate the use of AIS by fishing vessels, it may provide a useful source of information to map fishing activities stated that the data are interpreted in an appropriate way.
Improved understanding of human-nature interactions is crucial to conservation science and practice, but collecting relevant data remains challenging. Recently, social media have become an increasingly important source of information on human-nature interactions. However, the use of advanced methods for analysing social media is still limited, and social media data are not used to their full potential. In this article, we present available sources of social media data and approaches to mining and analysing these data for conservation science. Specifically, we (i) describe what kind of relevant information can be retrieved from social media platforms, (ii) provide a detailed overview of advanced methods for spatio-temporal, content and network analyses, (iii) exemplify the potential of these approaches for real-world conservation challenges, and (iv) discuss the limitations of social media data analysis in conservation science. Combined with other data sources and carefully considering the biases and ethical issues, social media data can provide a complementary and cost-efficient information source for addressing the grand challenges of biodiversity conservation in the Anthropocene epoch.