Опубликована: Янв. 1, 2024
Язык: Английский
Scientific Data, Год журнала: 2025, Номер 12(1)
Опубликована: Фев. 7, 2025
Advances in underwater imaging enable collection of extensive seafloor image datasets necessary for monitoring important benthic ecosystems. The ability to collect imagery has outpaced our capacity analyze it, hindering mobilization this crucial environmental information. Machine learning approaches provide opportunities increase the efficiency with which is analyzed, yet large and consistent support development such are scarce. Here we present BenthicNet: a global compilation designed training evaluation large-scale recognition models. An initial set over 11.4 million images was collected curated represent diversity environments using representative subset 1.3 images. These accompanied by 3.1 annotations translated CATAMI scheme, span 190,000 A deep model trained on preliminary results suggest it utility automating small-scale analysis tasks. made openly available reuse.
Язык: Английский
Процитировано
1
Geosciences, Год журнала: 2025, Номер 15(1), С. 7 - 7
Опубликована: Янв. 1, 2025
Initiatives such as the United Nations Decade of Ocean Science for Sustainable Development and Seabed 2030 promote seabed mapping worldwide. In Brazil, especially on Espírito Santo Continental Shelf, high-resolution has revealed an unknown complex seascape. Circular depressions (CDs) were mapped first time in Costa das Algas Marine Protection Area. Herein, we aim to present CD metrics characteristics discuss their relationship with morphology relevance a habitat based multibeam bathymetry ground truthing. A total 3660 between 46 85 m depth, reaching area 460 m2 5 relief. The continental shelf was subdivided into three sectors morphology: inter-valleys, valley edges, flanks, eleven sites selected direct sampling/imaging at CDs along sectors. sampling carried out by scuba-diving video images sediment samples collected inside outside depressions. deeper central parts circular appear function sink, presenting aggregations rhodoliths or other carbonate fragments. most inter-valley depressions, mounds observed edges depression. We did not have any indication gas seeps no clear sedimentological morphological control occurrence. hypothesize that origin results from combined diachronous processes. high resolution could be related sea level processes acting during last glacial period exposure, i.e., relict features. are responsible biomass aggregation fish bioturbation, forming holes rubble mounds, representing modern process occurring centimetric scale. data so far indicate this fine-scale feature is important different species. maintenance these structures due low sedimentation regime areas shaped biotic excavation.
Язык: Английский
Процитировано
0
Estuarine Coastal and Shelf Science, Год журнала: 2025, Номер unknown, С. 109203 - 109203
Опубликована: Фев. 1, 2025
Язык: Английский
Процитировано
0
Frontiers in Marine Science, Год журнала: 2025, Номер 12
Опубликована: Май 22, 2025
A spatially explicit understanding of marine benthic habitats is essential for sustainable resource management. While advances in remote sensing, acoustic methodologies, geostatistical modelling, and predictive species distribution models have improved our ability to map underwater habitats, selecting the most appropriate approach, particularly turbid or regions, remains challenging. This study was conducted protected nursery area Exmouth Gulf Prawn Managed Fishery Western Australia compared four commonly used “off-the-shelf” mapping techniques. These included satellite sounding, interpolation, with each technique evaluated using comprehensive ground-truthing output confidence matrices. Geostatistical kriging emerged as robust method, delivering highest accuracy, quantifiable confidence, seasonal habitat maps. maps delineated submerged aquatic vegetation, including seagrass macroalgae, at broad spatial scales captured shifts density. Our findings enhance knowledge underscore that effective mapping, dynamic environments, cannot rely on methods alone. Spatially balanced field data collection ecologically relevant temporal support
Язык: Английский
Процитировано
0
Earth Science Informatics, Год журнала: 2025, Номер 18(3)
Опубликована: Июнь 3, 2025
Язык: Английский
Процитировано
0
Frontiers in Remote Sensing, Год журнала: 2024, Номер 5
Опубликована: Ноя. 19, 2024
In recent decades, the value and utility of multibeam bathymetry backscatter data has been increasingly recognized within field seascape ecology (the ocean-centric equivalent landscape ecology) to map bio- geodiversity ocean floor. More recently, multispectral emerged as a promising endeavor in seabed classification, its acoustic response across multiple frequencies studied range substrates. Coral reef systems are an under-represented type this research, these shallow water ecosystems more commonly mapped with optical remote sensing techniques. Further investigation is necessary determine extent which can contribute characterization habitats–especially for that beyond reach approaches. This study investigates multifrequency coral reefs two areas on Abrolhos Shelf, Brazil. Backscatter mosaics angular curves (ARC) investigated explore potential applications enhanced characterization. both survey areas, mapping revealed extensive features, assessment was influenced by surface roughness biological attributes (algal cover). Results significantly understanding multi-frequency signatures seascapes highlight monitoring health ecosystems. Herein, higher were found be fundamental detection subtle variations texture inter-reef region areas; underscores employing framework discerning heterogeneities coarse substrate types. Furthermore, lower indispensable precise delineation characteristics Channel area, whereas most effective identifying epibiotic cover (those characterized flatter ARC associated values) Forgotten Reefs region.
Язык: Английский
Процитировано
2
Scientific Reports, Год журнала: 2023, Номер 13(1)
Опубликована: Ноя. 18, 2023
Abstract Acoustic backscatter has been used as a tool to map the seafloor in greater detail and plays an increasingly important role mapping meet multiple ocean management needs. An outstanding challenge use of for is distinction between acoustically similar substrates, such mixed sediments from rhodoliths. Rhodolith beds are biogenic substrate that provides ecological services, typically classified single categorical type—though nodules coverage may be spatially variable. Recently, multispectral acoustic demonstrated great potential improve thematic compared single-frequency systems. This work employs multibeam underwater imagery characterize rhodolith Costa das Algas Marine Protected Area (Brazil). A support vector machine classifier was classify multifrequency mosaics according classes identified imagery. Results suggest effective both providing information different proportions predicting presence or absence these nodules. The lowest frequency most useful distinguishing variable coverage, two higher frequencies were better predictors absence.
Язык: Английский
Процитировано
5
Marine Geology, Год журнала: 2024, Номер 469, С. 107243 - 107243
Опубликована: Фев. 15, 2024
Язык: Английский
Процитировано
1
Sustainability, Год журнала: 2024, Номер 16(6), С. 2374 - 2374
Опубликована: Март 13, 2024
Habitat mapping plays a crucial role in assessing marine protected areas (MPA) and implementing spatial management approaches. This study aims to present the habitat distribution of Jabal Ali Marine Sanctuary, considering development projects implemented its proximity. It serves as reference for guiding conservation efforts. The focuses on situ hyperspectral measurements optical properties both water column substrate. Additionally, high density geo-referenced spot checks were conducted, serving sample points ecological evaluation ground-truthing. An “object-oriented” approach was adopted generate seabed map two evaluated studies conducted 2006 2017. While 2017 survey identified 16 habitats, characterized only 10 habitats. These maps serve powerful tools mitigation measures providing scientific support mitigate negative impact most habitats within context area framework. Furthermore, monitoring cover important provides an integrative indicator maintain good environmental status sanctuary. Based this study, information will be evaluating synergizing approaches by competent authority different stakeholders
Язык: Английский
Процитировано
1
Опубликована: Янв. 1, 2024
Hydroacoustic sensors are better suited for mapping shallow turbid coastal waters, but surveys must take place along transects, necessitating the interpolation of processed, large, spatially dependent hydroacoustic datasets to produce continuous maps. The aim this study was evaluate performance techniques that can handle data obtained from a survey (covering 23 km2) conducted in 2016 bay southwestern Jamaica. Independent or 5-fold cross validation were used compare accuracies maps benthic sediment percentage composition and submerged aquatic vegetation cover (SAV), interpolated using random forest (RF) regression (RFr), RFr kriging (RFrk) Ensemble Machine Learning (EML) with rotated/oblique coordinates. Categorical/thematic types RF (RFc) EML classifiers spatial Markov chain model indicator predictor/simulation (Mc-IK) also compared. RFr, RFc, RFrk covariates included coordinates auxiliary data. following more accurate than corresponding maps: - mud, sand, silt, clay SAV; loss on ignition [LOI]; calcium carbonate; Mc-IK – 3 4 classes. most suitable assessing quantitative changes. predominant substrate type muddy sand covering 15.7 km2. highest clay, organic matter [LOI] found close mouth Black River where sandy mud found, indicating high input terrigenous sediment. Calcium carbonate showed an opposite trend positively associated higher SAV cover. is now part protected area, as such, these will be baseline gauge impact increased protection/management dynamics.
Язык: Английский
Процитировано
0