Echosounder Research Articles

Subsidence of Artificial Reefs with Bamboo Foundations on a Soft Seabed

The design of artificial reefs must be based on the criteria of both durability and serviceability. In this respect, the seabed subsidence rate related to the weight of the reefs and of their foundation system should be limited. The horizontal drift of the reefs during their life should be small too. In this study, pertaining to the behavior of artificial reefs made of microsilica concrete and realized with foundations in bamboo near the international airport of Hong Kong, 1-year monitoring data and calculations of reef foundation settlements are shown and discussed to assess the performance of the adopted mixed foundation and to deduce the nature of the long-term subsidence. The subsidence related to the settlements of the foundations was monitored using a Multibeam Echo Sounder System. It was found that the observed subsidence rate cannot be justified by the primary consolidation of the soils forming the seabed. Indeed, secondary consolidation seems to play an important role. Overall, after one year of service, the artificial reefs are considered to be effective in terms of limitation of both subsidence and drift.

Impact of Typhoon Nanmadol (2011) on the propagation of nonlinear internal waves in the South China Sea

In situ measurements from an array of pressure-equipped inverted echo sounders (PIESs), numerical simulation results using a non-hydrostatic model (SUNTANS) that does not account for mesoscale variability, and a ray-tracing method are used to investigate the behavior of nonlinear internal waves (NLIWs) in the South China Sea, focusing on period of Typhoon Nanmadol passage. Unusual significant differences in NLIWs arrival time occur between the PIES observation and SUNTANS simulation results after Typhoon Nanmadol passage; i.e., the observed NLIWs show a delay of 0.5–1.5 h at the westernmost PIES sites compared to the simulated ones in the SUNTANS. A data-assimilated ocean model outputs in addition to satellite altimetry reveal that the passage of Typhoon Nanmadol alters the oceanic environments remarkably, creating thermocline shoaling of ~80 m, which can slow down the wave propagation speed. We account for the delay of NLIWs using ray-tracing simulations. Our results demonstrate that typhoon-induced changes in mesoscale structure can significantly impact the fate of NLIWs in the South China Sea, which have potential on ocean mixing and biogeochemical systems.

Оценка фактического объема воды в водоеме посредством надводного беспилотного аппарата

Purpose: is to assess the actual water volumes in water objects (reservoirs, lakes, ponds) to identify the degree of maximum anthropogenic load, namely, the pollutant release. Materials and methods. Lake Svyatoe, located in Kosino-Ukhtomsky administrative district of Moscow, was chosen as the research and experimental object for testing the automatized system. The programs Google Earth Pro, Surfer 22, local author's programs for data processing, author's unmanned surface vehicle (USV), echo sounder Garmin Striker Cast GPS were used in the study. Automated systems as an unmanned surface vehicle (unmanned boat – USV), sensors (including echo sounder) and local programs for data acceptance and processing can be applied to improve and organize continuous monitoring of water objects. Results. The actual water volume assessment in the water object of 215 thousand cubic meters by means of the USV and its mounted echo sounder was demonstrated. The methodology of the unmanned boat operation for the water resources exploration was developed. Conclusions. Taking Lake Svyatoe as an example of water resources volume survey, it can be concluded that the methodology is applicable to any other water body, including reservoirs. Through regular water resources survey using USV it is possible to record not only changes in the actual water resources volume, but also bottom relief changes, sedimentation or overgrowing processes at specific sites. It also allows calculating the maximum allowable anthropogenic impact on the given object, including the maximum allowable discharge and its intensity at a particular moment of time.

Morphometric and hydrological features of the lakes of the Sergachsky municipal district (Nizhny Novgorod region, Russia)

The article presents the results of a study of the morphometric and hydrological features of the lakes of the Sergachsky municipal district (Nizhny Novgorod region, Russia), lying in the north of the Privolzhskaya upland. A bathymetric survey was carried out on the lakes of the researched territory from the board of an inflatable boat, using an echo sounder, combined with a navigator. The results of the field research were processed using the geographic information program QGIS, digital models of the relief of the six researched lakes were created, and their main morphometric and hydrological characteristics were calculated. Based on the analysis of the morphometric characteristics of the lakes and data on the landscape features of their surroundings, conclusions were drawn about the origin of the lake basins. Two genetic groups of lakes within the researched area are identified: karst and riverbed (oxbow lake).

Modelling of Kuibyshev reservoir shallow water depths by bathymetric surveys and multispectral UAV imagery data: A case study

This study investigates the potential of using multispectral Unmanned Aerial Vehicle (UAV) imagery to model the shallow water depths of the Kuibyshev Reservoir, Russia. Traditional methods like boom soundings and echo sounders, while accurate, are labor-intensive and costly. By leveraging multispectral data from UAVs, we aim to provide a more efficient and detailed approach to bathymetric mapping. Our methodology involved conducting bathymetric surveys with a Garmin GPS Map 178C and a Geoscan 401 Geodesy UAV equipped with a MicaSense RedEdge-MX camera. We performed correlation analysis and modelled depth using various regression techniques, identifying the Decision Tree Regressor as the top-performing model with an R² value of 0.98. Our findings suggest that UAV multispectral bathymetry is a viable alternative for local-scale shallow water mapping, with significant implications for reservoir management and ecological studies.

Regional Abyssal Vorticity Balance in the Northeast South China Sea: External and Internal Dynamics of Abyssal Circulation

Abstract Abyssal vorticity balance in the northeast South China Sea was assessed for over a year based on observations from 28 current- and pressure-recording inverted echo sounders distributed west of the Luzon Strait. The regional first-order balance was dominated by the planetary vorticity flux and bottom pressure torque, which reflect the external and internal dynamics of abyssal circulation. Vertical motion considerably contributed to the planetary vorticity flux, whereas the contribution of horizontal motion was negligible. Positive and negative planetary vorticity fluxes dominate the areas along the eastern and western boundaries, indicating upward and downward vertical transport, respectively. The opposite planetary vorticity fluxes in the different areas were accompanied by different current patterns; regional anticyclonic and cyclonic characteristics appeared near the western and eastern boundaries, respectively, owing to the deep topography as the abyssal current followed the boundary. The planetary vorticity flux near the eastern boundary was substantial in spring and autumn; in contrast, along the western boundary it was enhanced only in spring. Deep eddies played important roles in planetary vorticity flux and regional vorticity balance. The results of this study reveal the formation dynamics of abyssal circulation in the South China Sea as well as its spatiotemporal distributions, providing a more detailed description of abyssal circulation. Significance Statement The deep South China Sea (SCS) is a nearly enclosed basin characterized by cyclonic abyssal circulation. Based on the observations from 28 current- and pressure-recording inverted echo sounders distributed west of the Luzon Strait, the vorticity balance in the deep SCS was clarified. The planetary vorticity flux and bottom pressure torque maintain a first-order balance of vorticity, which acts as the external and internal dynamics of the abyssal circulation. The study describes the temporal variability and spatial distribution of vorticity terms in the deep ocean west of the Luzon Strait, which may contribute to a more detailed understanding of abyssal circulation formation and its evolution.

Hidrografski premer međunarodnih plovnih puteva u Republici Srbiji

European inland waterways are a key element of the European transport network and are mostly used for freight traffic. The hydrographic surveying is the basis for the maintenance of inland waterways and the establishment of safe navigation. A part of the European waterway corridor E80 consists of the rivers Danube, Sava and Tisza, which part of their course also flows through the Republic of Serbia. On these sections of the mentioned rivers, the competent state institutions are conducting a continuous campaign of hydrographic measurements, which, among other things, includes GNSS RTK positioning and depth measurement with a single-beam ultrasonic echo sounder. In this paper, the characteristics of these sections of rivers, the current methodology of depth measuring and positioning during depth measuring will be shown in details.

Active Pose - Autonomous Hole-Filling Techniques for 3D Surface Reconstruction in Underwater Exploration

Underwater exploration and related research face significant challenges, including incomplete data collections and the high cost of operations. An underwater robot's capacity to independently identify, target, and address gaps in 3D surface data, termed as ‘holes’, during a mission has the potential to significantly lower the expenses associated with research endeavors, such as those in underwater archaeology or subsea monitoring, and to improve sea surface mapping for autonomous robots. This study introduces a novel approach that combines the boundary-edge of a triangle mesh with Principal Component Analysis (PCA) to actively gather incomplete data. We propose two distinct techniques for active hole-filling, as well as a method to determine the most appropriate strategy. We introduce the flying-over and following-rail approaches to automatically determine new sensor positions for collecting additional point cloud data, which are then used to reconstruct the surface and fill the gaps. Experiments were conducted using both simulated underwater images with ORB-SLAM3 (Campos et al. (2021)) and real-world Multibeam Echo Sounder data.

CAREER PROFILES • OPTIONS AND INSIGHTS: Scott Loranger

I received a Bachelor of Science in biological sciences with a focus on ecology and evolutionary biology from Cornell University in 2009. I also spent a semester with SEA Education Association out of Woods Hole, Massachusetts, and that experience solidified my interest in oceanography. As I delved deeper into marine science, I became interested in ecological resilience and response to change. Following my undergraduate education, I worked with a group at the University of Maryland Center for Environmental Science – Chesapeake Biological Labora­tory studying technology to treat ballast water to deter transport of invasive species, and technology to measure oil concentrations in the ocean to respond to oil spills. This was my first foray into applied science and emerging technologies. My interest in technology and oil spill response led me to the University of New Hampshire (UNH) and the Center for Coastal and Ocean Mapping, where I obtained a PhD in oceanography, graduating in 2019. At UNH, I studied the use of broadband acoustic echo sounders to map, characterize, and quantify oil spills. For my thesis, I studied the 2004 Taylor Energy oil spill in the Gulf of Mexico, the longest-running oil spill in US history.

Dense Neural Network for Classification of Seafloor Sediment using Backscatter Mosaic Feature

Water transportation plays a vital role in global economic activities, facilitating more than 85% of international trade and serving as a cost-effective and essential means to fulfill the demand for goods and services. Similarly, the Benoa Port, situated in the southern part of Denpasar City, operates in the same manner. By utilizing Multibeam Echo Sounder (MBES) backscatter data, backscatter mosaics can be generated to identify various seafloor sediment types, which consist of rock fragments, minerals, and organic materials. The characteristics of these sediments, such as grain size, density, composition, and others, can be observed. To improve the classification of sediments, the integration of backscatter data and backscatter features, such as ASM (Angular Second Moment), Energy, Contrast, and Correlation, can be employed. Supervised classification models like Dense Neural Network (DNN) can be utilized to accurately determine the types of seafloor sediments. The application of DNN modeling resulted in a training accuracy rate of 88% and a testing accuracy rate of 100%. The accuracy results delineated six distinct sediment types. Notably, sandy silt exhibited the highest distribution, accounting for 49.30%, whereas soft clayey silt registered the lowest distribution at 0.53%, as determined by their respective spatial prevalence.

A contour-based path planning method for terrain-aided navigation systems with a single beam echo sounder

Underwater terrain-aided navigation (TAN) methods with a single beam echo sounder (SBES) have the potential to facilitate long-term underwater navigation for underwater vehicles. However, the limited bathymetric measurement of the SBES often makes the accuracy and robustness of TAN navigational results greatly affected by the complexity of the terrain. Therefore, effective path planning for SBES-TAN is essential. This paper proposes a contour-based path planning method for SBES-TAN, which can yield robust navigational results even with noisy SBES measurements. Moreover, both the density and smoothness of contour lines are used to quantify terrain complexity. Moreover, a fuzzy comprehensive evaluation-based theory is used to identify TAN-suitable areas. Path planning is accomplished using the rapid-exploration random tree* algorithm, while a Monte Carlo-based TAN estimate handles the challenges posed by noisy SBES measurements. The performance of the proposed method has been verified through simulation using field bathymetric data, by comparing it with state-of-the-art methods.

Detailed Bathymetry and Seafloor Backscatter Image Dataset for Monitoring Ecosystem Environment: Southern Ulleungdo

We obtained a detailed bathymetry dataset using multi-beam echo sounder around Neunggeol located at Southern part in Ulleungdo. The survey period was from September 1, 2019 and the bathymetry survey was carried out by RV Jangmok No2 of Korea Institute of Ocean Science and Technology. The instrument equipment used for the survey were Kongsbergs’s EM2040 (multi-beam echo sounder). The detailed bathymetry of the study area shows gradually depth change from about 4 to 190 m. The study area has relatively gentle slopes from the coast to 25 m and below about 125 m in depth, but a steep slope from about 25 to 125 m. The rocky bottom zones identified from detailed bathymetry and backscattered seafloor images are distributed in the northeastern sea of the coastal area, the central sea, and around Neunggeol. In particular, Neunggeol and the near large outcropping rock are tumulus-like submarine landforms, forming big vertical walls. The submarine cable buried between Ulleungdo and the Korean mainland appears in the northern seafloor of the survey area. Stopper stones and gravels sediment are distributed around it. The rocky bottom zone has a strong and irregular reflection intensity in the seafloor backscattered image and the sandy sediment zone has a relatively weak reflection intensity and a uniform pattern, so the two zones are well divided. We classified the sediment zones into fine-grained sediments, sandy sediments, and gravel-like sediments by using the different reflection intensities due to differences in sediment particle size in the seafloor image. Detailed bathymetry and seafloor backscattered image dataset around Neunggeol can be used as a basic data for habitat environment mapping.

Archaeological Exploration, Excavation, and Analysis for a Richer Interpretation of the Past Chola Heritage City- Poompuhar

The article deals with the process of collecting underwater data from the historical area in South India and explicitly showcasing its ancient civilization. Various methods are used to investigate the environmental and structural variables that regulate the coastal structures. Geostatistical interpolation is a technique to predict the values of a spatially continuous bathymetry or depth of a submerged shoreline from a few sample data measurements. The bathymetric investigation of the submerged region surrounding the ancient port of Poompuhar consists of numerous processes conducted in stages. The survey was carried out with the use of an integrated measuring system developed by the National Institute of Technology, Chennai, which included various levels of Multi Beam Echo Sounder (MBES), GPS, sonar, and ROV. To discover continuous surfaces required for analyzing the morphology of the bottom of submerged Poompuhar, a suitable interpolation technique must be used to get estimated values in regions that were not physically surveyed. Bathymetric and topographic data, which typically are gathered independently for various purposes, were included in the spatial data used for elevation surface modeling. Data is gathered in multiple forms with varying resolutions and accuracy; as a result, a standard surface model that will allow for quick and accurate analysis is currently lacking. The major purpose of this study was to create a high-accuracy model of a coastal area's surface using input data from numerous sources. ArcGIS is a popular software platform for analyzing and visualizing geographic data, and it contains several geostatistical interpolation techniques. The research involves the erosion of the shoreline as well as the rise in water level. The study uncovered new scientific and methodological information on Poompuhar's bathymetric properties and submerged surface.

A method for extracting the spatial distribution of submerged vegetation in lakes based on topography measurements

Aquatic vegetation in lakes provides various ecosystem services, identification of its spatial distribution has great importance in lake ecosystem management and conservation. Monitoring of submerged aquatic vegetation is limited due to plants being under the water surface. Based on topography measurements, a new method was proposed to identify the distribution of submerged vegetation. With submerged vegetation occurred in lakes, water depth measured by an echo sounder could be influenced by vegetation and differ from its true value. Vegetation distribution can be extracted by comparing the topographic maps measured with and without vegetation presence. In present study, identification for Potamogeton crispus distribution in Dongping Lake (China) was taken as an example to show the procedures. Errors and accuracy of this method were analyzed, and a discussion was provided about the usefulness and limitations of the proposed method. In this study, measurements were conducted in March (with P. crispus occurrence) and September (without P. crispus occurrence), 2021. Five interpolation methods with ten resampling intervals were used and compared for each measurement, and the minimal mean absolute errors (MAEs) for interpolated topographies in Mar. and Sep. measurements were 0.63 m and 0.46 m, respectively. MAE for the subtraction of raster data between two topographic maps over the entire lake was 0.59 m. Besides, accuracy of this method was validated by comparing the extracted distribution of P. crispus with its distribution of biomass. Taking the P. crispus distribution with dry biomass density larger than 26 g/m2 as reference, the identification accuracy of this method reached its maximum (78%). In regions with P. crispus sparsely distributed, the measured water depth was seldom influenced by vegetation, leading to the differences between two maps ignorable in these regions.

Use of visual and echo sounding surveys: case study of submerged aquatic vegetation of lake Kenon (Eastern Transbaikalia)

The paper presents the results of studying of the aquatic vegetation of Lake Kenon based on the integrated use of traditional methods of visual survey with an echo sounding. The use of echo sounding surveys makes it possible to study better the features of the distribution of submerged aquatic plants. It allows you to speed up the process of sampling in the field, and also facilitates office processing of the received data. Based on the measurement results, a bathymetric map was built, and the boundaries and areas of dominant phytocenoses were determined. It was revealed that at present the total area of thickets is 489 hectares, which is 33 % of the lake area. The maximum depth of plant growth was 5.1 m. The data obtained is higher than the percentage of overgrowth in 1993, but lower than the values for the 70s and 80s. At present, as in previous years, Chara algae are leading in terms of overgrowing area. The decrease in the distribution area of Nitella flexilis and Potamogeton crispus was revealed.

Deep Neural Network to Classify Seabed Sediment Using MBES Multifrequency

This study investigates the classification of sediment distribution and types in inland waters using Artificial Neural Network (ANN) models based on multibeam echosounder data. Multibeam Echosounder Technology (MBES) is an acoustic instrument that emits high-speed sound pulses to map the bottom of water bodies. Valuable information about sediment characteristics can be obtained by analyzing the backscattering phenomenon, which measures the acoustic intensity reflected from the water bottom. The angular response curve (ARC) method, developed by Fonseca and Calder, is employed to analyze the relationship between backscatter intensity and the angle of reflection. The ANN models in this study are trained and tested using intensity values and angular response curves obtained from multibeam echo sounders. The ANN model architecture consists of an input layer, three hidden layers, and an output layer. The sigmoid activation function is utilized in the output layer, while the ReLU activation function is applied to the input and hidden layers. Performance evaluation demonstrates that the ANN model using uncorrected data achieves remarkable accuracy, with 100% accuracy obtained in both the training and testing stages. The model exhibits a training loss value of 0.059443 and a testing loss value of 0.110030, indicating its effectiveness in capturing patterns and reducing errors. Visualizing the sediment distribution in Waduk Selorejo reveals a relatively balanced presence of Clavey sand and Sandy silt sediment types. However, noticeable differences arise between the corrected and uncorrected data, which can be attributed to the alterations made during the data correction process. The best-performing ANN model achieves a testing accuracy of 0.75, indicating that it does not perfectly predict the given data. Acknowledging the potential for prediction errors when using corrected data is important.

Propagation Features of Diurnal Internal Tides West of the Luzon Strait Revealed by a Large PIES Array

Abstract Energetic internal tides (ITs) are generated from the Luzon Strait (LS) and propagate westward into the South China Sea (SCS). Owing to the lack of large-scale synchronous measurements, the propagation features and seasonal variations of diurnal ITs remain unclear. From 2018 to 2019, mode-1 diurnal ITs west of the LS were continuously observed using a large-scale moored array of 27 pressure-recording inverted echo sounders (PIESs) and a thermistor chain. Measurements confirmed that diurnal ITs radiate from the LS with a north–south asymmetrical pattern, with the most energetic channel located in the middle and south of the LS. The total energy radiated into the SCS across 120°E is 2.67 GW for the K1 ITs and 1.54 GW for the O1 ITs, approximately 2 times larger than those inferred from satellite observations. K1 dominates among the diurnal ITs, with its maximum isopycnal displacement (amplitude) and energy input to the SCS being the strongest in summer (i.e., 16.3 m and 2.81 GW, respectively). The propagation speed of K1 is higher in summer and autumn along the main channel (i.e., 4.33and 4.36 m s−1, respectively). Seasonal stratification and circulation play important roles in the seasonal variation of amplitude and propagation speed of the K1 ITs. The seasonal variability of diurnal-band ITs, which includes all diurnal constituents, is location-dependent and primarily results from the superposition of the K1 and P1 ITs. In particular, vertical displacement is strong in summer and winter along the main channel of the K1 and P1 ITs. The seasonal amplitude of K1 can modulate this seasonal feature. Significance Statement Internal tides (ITs) are internal waves at tidal frequencies. The Luzon Strait (LS) is one of the most energetic sites to generate large-amplitude ITs. The ITs propagate into the South China Sea (SCS), interact with mesoscale eddies, large-scale circulations, etc., and influence local hydrodynamics as well as ecosystem and sediment transport. This motivated an observation plan to investigate the ITs at the western entrance of the LS. From June 2018 to August 2019, an array of 28 PIESs was deployed in the northeastern SCS, almost covering the western entrance of the LS, to investigate the propagation properties of ITs including their amplitude, phase speed, wavelength, propagation direction, and energy fluxes and their annual and seasonal variations. Here, we primarily focus on the mode-1 diurnal ITs. The new insights enrich our understanding of IT dynamics and seasonal variations and support further improvements in numerical simulations.

Recent increase in sediment dry matter, carbon, and phosphorus accumulation in small boreal lakes with clayey catchments

This study estimated the mass accumulation rates of sediment (MAR), carbon (CMAR), and phosphorus (PMAR) in small Finnish lakes with agricultural clayey catchments over a 25-year period (1986–2011) and compared these with the conditions before major agricultural land use. Twenty-two lakes were cored for recent and reference (pre-disturbance) sediments. The recent sediment section was selected based on the 1986 137Cs fallout peak (TOP), whereas the pre-disturbance section (REF) was selected immediately below the first signs of human-induced erosion. The 50-cm reference section was dated with 14C at both ends. The mass accumulation rates were estimated based on dating, weighing, and chemical analysis for both sediment sections. Furthermore, sediment-penetrating echo soundings were used to estimate the amount of sediment in the whole lake basins. These data were used to examine area-specific loading from clay-rich catchments. The average whole-basin pre-disturbance MAR, PMAR, and CMAR were 62 g m−2 a−1, 0.06 g m−2 a−1, and 4.7 g m−2 a−1, respectively. The corresponding recent rates were 11, 13, and 8-fold (693 g m−2 a−1, 0.79 g m−2 a−1, and 37 g m−2 a−1). In the recent conditions, sediments were generally more minerogenic and MAR, PMAR, and CMAR were higher in lakes with more arable fields in their catchments. Average area-specific suspended sediment load from the catchment for the region (∼39% clayey soils) was approximately 69–137 kg ha−1 a−1 in the undisturbed state and 767–1534 kg ha−1 a−1 in recent conditions based on 100–50% retention. The results demonstrate that the increases in sediment, nutrient, and organic matter accumulation due to agriculture can be several fold over undisturbed state.

Anatomy of Mode-1 Internal Solitary Waves Derived From Seaglider Observations in the Northern South China Sea

Abstract Observations from a Seaglider, two pressure-sensor-equipped inverted echo sounders (PIESs), and a thermistor chain (T-chain) mooring were used to determine the waveform and timing of internal solitary waves (ISWs) over the continental slope east of Dongsha Atoll. The Korteweg–de Vries (KdV) and Dubreil–Jacotin–Long (DJL) equations supplemented the data from repeated profiling by the glider at a fixed position (depth ∼1017 m) during 19–24 May 2019. The glider-recorded pressure perturbations were used to compute the rarely measured vertical velocity (w) with a static glider flight model. After removing the internal tide–caused vertical velocity, the w of the eight mode-1 ISWs ranged from −0.35 to 0.36 m s−1 with an uncertainty of ±0.005 m s−1 due to turbulent oscillations and measurement error. The horizontal velocity profiles, wave speeds, and amplitudes of the eight ISWs were further derived from the KdV and DJL equations using the glider-observed w and potential density profiles. The mean speed of the corresponding ISW from the PIES deployed at ∼2000 m depth to the T-chain moored at 500 m depth and the 19°C isotherm displacement computed from the T-chain were used to validate the waveform derived from KdV and DJL. The validation suggests that the DJL equation provides reasonably representative wave speed and amplitude for the eight ISWs compared to the KdV equation. Stand-alone glider data provide near-real-time hydrography and vertical velocities for mode-1 ISWs and are useful for characterizing the anatomy of ISWs and validating numerical simulations of these waves. Significance Statement Internal solitary waves (ISWs), which vertically displace isotherms by approximately 100 m, considerably affect nutrient pumping, turbulent mixing, acoustic propagation, underwater navigation, bedform generation, and engineering structures in the ocean. A complete understanding of their anatomy and dynamics has many applications, such as predicting the timing and position of mode-1 ISWs and evaluating their environmental impacts. To improve our understanding of these waves and validate the two major theories based on the Korteweg–de Vries (KdV) and Dubreil–Jacotin–Long (DJL) equations, the hydrography data collected from stand-alone, real-time profiling of an autonomous underwater vehicle (Seaglider) have proven to be useful in determining the waveform of these transbasin ISWs in deep water. The solutions to the DJL equation show good agreement with the properties of mode-1 ISWs obtained from the rare in situ data, whereas the solutions to the KdV equation underestimate these properties. Seaglider observations also provide in situ data to evaluate the performance of numerical simulations and forecasting of ISWs in the northern South China Sea.

Addressing class imbalance in deep learning for acoustic target classification

Abstract Acoustic surveys provide important data for fisheries management. During the surveys, ship-mounted echo sounders send acoustic signals into the water and measure the strength of the reflection, so-called backscatter. Acoustic target classification (ATC) aims to identify backscatter signals by categorizing them into specific groups, e.g. sandeel, mackerel, and background (as bottom and plankton). Convolutional neural networks typically perform well for ATC but fail in cases where the background class is similar to the foreground class. In this study, we discuss how to address the challenge of class imbalance in the sampling of training and validation data for deep convolutional neural networks. The proposed strategy seeks to equally sample areas containing all different classes while prioritizing background data that have similar characteristics to the foreground class. We investigate the performance of the proposed sampling methodology for ATC using a previously published deep convolutional neural network architecture on sandeel data. Our results demonstrate that utilizing this approach enables accurate target classification even when dealing with imbalanced data. This is particularly relevant for pixel-wise semantic segmentation tasks conducted on extensive datasets. The proposed methodology utilizes state-of-the-art deep learning techniques and ensures a systematic approach to data balancing, avoiding ad hoc methods.