Multibeam Sonar Research Articles

Discovering the Bathylithology and Bioengineering Organisms of the Punta Coles Marine Natural Reserve, Moquegua, Peru

The Punta Coles Natural Reserve (PCNR), located on the southern coast of the province of Ilo, Moquegua Region, Peru, houses an area rich in marine biodiversity which is composed of natural banks of benthic invertebrates, which constitute the potential and dynamic fishery of commercial interest for artisanal fishermen that impact the area. To mitigate this impact, all stakeholders need to know, understand, value, and balance the dependence of benthic biodiversity with the ecosystem for its preservation and the promotion of sustainable fishing. Along these lines, the components studied are bathymetry, lithology, and the tapestry of bioengineers that make up the reserve’s seabed. Mapping of the complexity of the benthic habitat of the seabed was developed on the basis of an integrated analysis of multibeam bathymetric data, geoscientific information associated with the antecedents of benthic biodiversity of the PCNR, sediment maps of the seabed, subtidal sedimentological texture, and photographic records. The characterization and granulometry of the sediments that make up the PCNR were determined according to the Udden–Wentworth classification criterion (Φ). The utility of the GPSMAP 178C multibeam echosounder for the interpretation of seabed sediments stands out. The kriging interpolation method was used to extrapolate the characteristics of the benthic habitat, which is suitable for large areas of the sea. As a result, three habitats were derived and mapped, and three structures (bioengineers), namely, “subtidal kelps” (Lessonia trabeculata), “mussels” (Aulacomya atra), and “red sea squirt” (Pyura chilensis), were associated mainly with sedimentary rock habitats and mixed substrates. In the sand substrate, these algae were found to a lesser extent. The results show that this study is a conceptual model that summarizes the understanding of ecology and prospects for the sustainability of profitable, ecoefficient fishing activities in the Moquegua Region, Peru.

Unveiling Gas Seeps: A Dive into Water Column Data Analysis

Abstract Multibeam Echo Sounders (MBES) have significantly advanced underwater acoustic data analysis, allowing for the detection of various marine features. The enhanced capabilities of water column acoustic data analysis by MBES have played a crucial role in numerous research projects, including mapping methane gas flux and studying free gas release. This research aims to detect and quantify underwater gas seeps by leveraging the advanced imaging capabilities of MBES. The workflow involved displaying water column data, removing irrelevant echoes, applying amplitude threshold filtering, validating Local Intensity Maxima (LIM) values, exporting final coordinates, and calculating volumes using voxel-based methods. Results showed that higher frequencies (200 and 400 kHz) provided more detailed and accurate gas seep detection at 30 m depth, with average volumes of 25.972 m3 and 31.050 m3, respectively. However, at a depth of 60 m, the 100 kHz frequency was more effective, with an average volume of 28.324 m3. These findings underscore the importance of frequency selection in MBES surveys for accurate gas seep detection and quantification. This study provides valuable insights into underwater gas seeps, enhancing our understanding of their impacts on the marine environment and the global carbon cycle. This study supports SDG 14 by advancing marine ecosystems, including the management of marine resources.

3D modeling and identification of post-dredging sediment layer types using hydroacoustic instruments

Abstract Sedimentation in harbor areas is a serious problem because it can interfere with ship navigation activities due to silting depth. Overcoming the siltation problem requires dredging to transport most sediment on the bottom of the water. In 2023, dredging was carried out to overcome the siltation problem, where the dredged layer was a soft layer and a small part of the hard layer. This research focuses on identifying the type of bottom sediment layer in PT Gresik Jasatama after dredging, modeling the sub-bottom layer and its depth and in addition, analyzing the difference in depth before and after dredging and interpreting the Side Scan Sonar image. The method used is an acoustic method, which consists of various instruments used to perform the analysis. The sub-bottom Profiler uses the acoustic wave reflection principle to determine the sediment layer type. Multibeam Echosounder is used to analyze the difference in bathymetry depth in 2021 and 2023 while analyzing the type of layer based on the digitation of the Sub-bottom Profiler. An echo-logger is used to analyze the difference in depth generated by low and high frequencies. Side Scan Sonar is used to provide a visual appearance of the shape of the seabed or objects on the seabed, and Backscatter MBES is used to analyze the type of sediment at the bottom of the water. Overlay between the difference data of Echologger, Sub-bottom Profiler, and bathymetry in 2023, a hard layer in the study area indicates dredging. In addition, the bathymetry measurement in the research area in 2021 was 1-8 meters, while in 2023, it was 7-12 meters, and the depth difference between the two depths was 4 meters. Side Scan Sonar interpretation shows the hard layer and soft layer by the borehole retrieval point, and the results of the water bottom characteristics show sediments in the form of gravel, gravel silty sand, clay, and sandy clay.

Comparison of multitemporal multifrequency multibeam echosounder norbit iwbms: depth and sedimentation rate analysis of selorejo reservoir 2022-2023

Abstract Selorejo Reservoir is a reservoir constructed and located in Pandansari Village, Ngantang District, Malang Regency, East Java. The Indonesian Government built this reservoir to accommodate the water flow from several rivers, including the Konto River, Kletak Lahar, Kwayangan River, and several other small rivers. The biggest challenge for a reservoir, where the primary clean water source comes directly from rivers, is siltation caused by sediment accumulation at the bottom of the water body. Therefore, severe and continuous control in the reservoir area is necessary to prevent damage and overflow and ensure the community’s safety by engaging in activities around the Selorejo Reservoir. Bathymetric surveys are essential in obtaining information related to marine fields and precise depth values with spatial references. When using a multibeam echosounder, surveyors must consider several aspects regarding the specifications of the equipment before data acquisition, one of which is the frequency used. Different frequencies result in different penetration abilities and data quality. This study aims to estimate the sedimentation rate in the Selorejo Reservoir using multitemporal data and investigate the effect of using multifrequency on the depth measurement results obtained. The results of this study indicate that the higher the frequency used, the shallower the depth value obtained. Higher frequencies provide better representation in object detection capabilities. At the same time, the sedimentation rate was relatively high, from 3.2 to 5.3 centimetres per month.

Comprehend Analysis of Surface and Subsurface Sediment Distribution Using Underwater Acoustic Instruments

Abstract Ports play an important role in shipping and transporting while its fundamental challenge is sedimentation that occurs continuously at a location. A management system is required to respond to this siltation problem by dredging the underwater sediments parallel with Sustainable Development Goal 11 about sustainable cities and communities. Therefore, analyzing the area’s sediment types and distribution is necessary, especially using acoustic instruments. In this research, the dredging plan shows the relative depth distribution is around 4 to 18 meters with deep sloping topography on the north side. Furthermore, the overall result of the backscatter value of the multibeam echosounder ranges from −45.86 dB to −55.73 dB. According to the Angular Response Analysis, the sediment type of the seabed surface is clay with depths ranging up to 5.40 meters. The result of the sub-bottom profiler processing discovers step or bowl-like features where the curvature seems to have an unequal height. It considers the difference in sediment height around the area while the intensity is relatively low. Through acoustic processing analysis, it informs that the total area of 643872 m2 needs to be dredged with an estimated dredging volume of around 2181105 m3 both in surface and subsurface areas. The qualitative analysis of Multibeam Echosounder, and Sub Bottom Profiler data shows that the acoustic instrument’s products are accurate with validation data in the field.

Advanced Detection of Underwater Gas Seep Sites Through Multibeam Echosounder Water Column Data and Numerical Analysis

Abstract Water column imaging of multibeam echosounder systems (MBES) are sensitive and potential devices for investigating free gas bubbles release and their ascent up the water column. These data could demonstrate previously undetected characteristics on the water’s surface and seabed which are related to Sustainable Development Goal 14 about life below water. The research utilizes a MBES to map the volume of seabed gas emission bubbles in the Adriatic Sea, Italy, using water column data. The survey covered 1.5 km2 around a four-legged gas platform at a depth of 77 meters. To achieve a 50% overlap, ten parallel transects of 1.5 km each were used, with a vessel speed range of 2-2.6 m/s and a transect spacing of 100 m. Acoustic waves from seabed seepage were visualized using water column data, with reflection intensities ranging from −63.5 dB to 29 dB, reflecting the acoustic reflectance of various suspended materials. More precise thresholds were obtained by filtering and clustering the gas bubbles using the point weight approach to separate them from the noise and water bubbles. The uneven Digital Terrain Model (DTM) indicates gas emissions through water column data. The volume of gas bubbles was determined by visualizing the points in a 3D format using XYZ coordinates. Through interpolation techniques and 3D volumetric analysis, six bubble locations were obtained with volumes of 651.12 m3, 108.30 m3, 42.00 m3, 167.20 m3, 186.00 m3, 287.81 m3, and 45.00 m3. This study is crucial because it questions the methods of Carbon Capture and Storage (CCS) by investigating the discharge of carbon into the sea. Furthermore, this study helps to identify emission sources, measure the volume of released gas, and explain the depth distribution of emissions, providing essential data for marine CCS assessment.

Ship heave measurement method based on sliding adaptive delay-free complementary band-pass filter

Accurate measurement of heave motion plays an important role in multi-beam echo sounding system and seabed feature recognition. Currently, scholars have put forth two primary methods for measuring ship heave, namely employing Kalman filtering and traditional high-pass filters. Nevertheless, the precision of heave measurement is somewhat compromised due to challenges in accurately establishing ship heave dynamics models, reliance on empirical values or peak detection ranges, and the impact of time-delay errors. To enhance the precision of heave motion measurement, this paper introduces a method based on a sliding adaptive delay-free complementary band-pass filter. The method adaptively acquires cutoff frequencies of band-stop filters within different windows and derives the transfer function of the non-delay complementary band-pass filter based on complementarity. To validate the efficacy of the algorithm, this study conducted experiments utilizing two distinct sets of shipborne data acquired on September 18, 2023, and September 22, 2023. The proposed method significantly improved accuracy compared to other methods, reducing reliance on prior knowledge or empirical values, and mitigating time-delay errors, thus achieving high-precision heave motion measurement.

Automated Shoreline Segmentation in Satellite Imagery Using USV Measurements

Generating aerial shoreline segmentation masks can be a daunting task, often requiring manual labeling or correction. This is further problematic because neural segmentation models require decent and abundant data for training, requiring even more manpower to automate the process. In this paper, we propose utilizing Unmanned Surface Vehicles (USVs) in an automated shoreline segmentation system on satellite imagery. The remotely controlled vessel first collects above- and underwater shoreline information using light detection and ranging (LiDAR) and multibeam echosounder (MBES) measuring instruments, resulting in a geo-referenced 3D point cloud. After cleaning and processing these data, the system integrates the projected map with an aerial image of the region. Based on the height values of the mapped points, the image is segmented. Finally, post-processing methods and the k-NN algorithm are introduced, resulting in a complete binary shoreline segmentation mask. The obtained data were used for training U-Net-type segmentation models with pre-trained backbones. The InceptionV3-based model achieved an accuracy of 96% and a dice coefficient score of 93%, demonstrating the effectiveness of the proposed system as a source of data acquisition for training deep neural networks.

Cross-PIC: A cross-scale in-context learning network for 3D multibeam point cloud segmentation of submarine pipelines

The multibeam sonar has become a powerful tool for undersea pipeline detection, which can obtain high quality three-dimensional point cloud data. However, the identification accuracy of current pipeline algorithms remains low because of insufficient utilization of context features. Therefore, this paper proposes an improved context-learning point cloud segmentation network, named Cross-scale Point-In-Context (Cross-PIC), to identify submarine pipelines. Cross-PIC designs the sampling module Balanced-FPS to improve the equalization sampling. Meanwhile, Cross-PIC constructs different scale coding architecture based on context learning, which includes Auxiliary line and Baseline for multi-scale feature encoding. Finally, the Refine Module is designed to realize cross-scale coded feature fusion to output segmentation results. Experiments with different fuzzy degrees show that Cross-PIC also significantly outperforms the original Point-In-Context network with less training data. Cross-PIC also achieves better results than other networks based on point, voxel, and dynamic graph convolution.

Depth gate tracking method based on historical sounding results in MBES

Gate tracking technology plays a crucial role in the autonomous detection of multi-beam echo sounding. The quality of the tracking algorithm directly impacts the quality of the estimate data. Here, we propose a new method for depth gate tracking in multi-beam echo sounding based on the assumption that historical sounding results are geometrically similar to the current ping results. Considering Hausdorff distance measures how far two subsets of a metric space are from each other. An evaluation method based on the concept of similarity distance that relies on the Modified Hausdorff distance is proposed here to measure the similarity between the detection results of different pings and determine the maximum number of historical pings that can be tracked. Then, this model is combined with the iterative closest point registration algorithm to align the bathymetric results of historical pings to the current ping, eliminating the need for prior registration. The registered data are used to initialize the particle distribution in the tracking particle filter and a distance-weighted importance function is established for each beam. Validation on measured data has shown that the proposed method is effective in tracking seabed topography and providing stable gate tracking results.

A Comparative Crash-Test of Manual and Semi-Automated Methods for Detecting Complex Submarine Morphologies

Multibeam echosounders provide ideal data for the semi-automated seabed feature extraction and accurate morphometric measurements. In this study, bathymetric and raw backscatter data were initially used to manually delimit the reef morphologies found in an insular semi-enclosed gulf in the northern Aegean Sea (Gera Gulf, Lesvos Island, Greece). The complexity of this environment makes it an ideal area to “crash test” (test to the limit) and compare the results of the delineation methods. A large number of (more than 7000) small but prominent reefs were detected, which made manual mapping extremely time-consuming. Three semi-automated tools were also employed to map the reefs: the Benthic Terrain Modeler (BTM), Confined Morphologies Mapping (CoMMa), and eCognition Multiresolution Segmentation. BTM did not function properly with irregular reef footprints, but by modifying both the bathymetry and slope, the outcome was improved, producing accurate results that appeared to exceed the accuracy of manual mapping. CoMMa, a new GIS morphometric toolbox, was a “one-stop shop” that, besides generating satisfactory reef delineation results (i.e., detecting the same total reef area as the manual method), was also used to extract the morphometric characteristics of the polygons resulting from all the methods. Lastly, the Multiresolution Segmentation also gave satisfactory results with the highest precision. To compare the final maps with the distribution of the reefs, mapcurves were created to estimate the goodness-of-fit (GOF) with the Precision, Recall, and F1 Scores producing values higher than 0.78, suggesting a good detection accuracy for the semi-automated methods. The analysis reveals that the semi-automated methods provided more efficient results in comparison with the time-consuming manual mapping. Overall, for this case study, the modification of the bathymetry and slope enabled the results’ accuracy to be further enhanced. This study asserts that the use of semi-automated mapping is an effective method for delineating the geomorphometry of intricate relief and serves as a powerful tool for habitat mapping and decision-making.

Seafloor sediment characterization improves estimates of organic carbon standing stocks: an example from the Eastern Shore Islands, Nova Scotia, Canada

Abstract. Continental shelf sediments contain some of the largest stocks of organic carbon (OC) on Earth and play a vital role in influencing the global carbon cycle. Quantifying how much OC is stored in shelf sediments and determining its residence time is key to assessing how the ocean carbon cycle will be altered by climate change and possibly human activities. Spatial variations in terrestrial carbon stocks are well studied and mapped at high resolutions, but our knowledge of the distribution of marine OC in different seafloor settings is still very limited, particularly in dynamic and spatially variable shelf environments. This lack of knowledge reduces our ability to understand and predict how much and for how long the ocean sequesters CO2. In this study, we use high-resolution multibeam echosounder (MBES) data from the Eastern Shore Islands offshore Nova Scotia (Canada), combined with OC measurements from discrete samples, to assess the distribution of OC content in seafloor sediments. We derive four different spatial estimates of organic carbon stock: (i) OC density estimates scaled to the entire study region assuming a homogenous seafloor, (ii) interpolation of OC density estimates using empirical Bayesian kriging, (iii) OC density estimates scaled to areas of soft substrate estimated using a high-resolution classified substrate map, and (iv) empirical Bayesian regression kriging of OC density within areas of estimated soft sediment only. These four distinct spatial models yielded dramatically different estimates of standing stock of OC in our study area of 223 km2: 80 901, 58 406, 16 437 and 6475 t of OC, respectively. Our study demonstrates that high-resolution mapping is critically important for improved estimates of OC stocks on continental shelves and for the identification of carbon hotspots that need to be considered in seabed management and climate mitigation strategies.

Benthic habitat mapping of the glass sponge (Vazella pourtalesii), and associated community composition on Sambro Bank, Scotian Shelf, Canada

Sponges have been identified as ecosystem engineers, providing habitat structure for many other benthic organisms, but are vulnerable to the impacts of bottom contact fishing. The Sambro Bank Sponge Conservation Area (SBSCA) protects a globally unique aggregation of the glass sponge (Vazella pourtalesii) off the east coast of Canada. Here, we present the first comprehensive, high-resolution mapping study of this site. Multispectral multibeam echosounder (MBES) mapping at three operating frequencies and benthic drop camera surveys were conducted at the SBSCA in 2022. Video data were analysed to quantify the abundance and location of V. pourtalesii and all associated benthic fauna and sediment characteristics. Using the MBES data as predictor variables, the study created a generalized linear mixed model of V. pourtalesii presence in the SBSCA, with 90.1% accuracy. Four seafloor benthoscape classes were identified from the video imagery, with statistically distinct benthic macrofaunal communities associated with each class. The study creates a baseline assessment of the current community composition and habitat characteristics of the site, which can be used for future ecological monitoring and management.

Atlas of pacific sand lance (Ammodytes personatus) benthic habitat – Application of multibeam acoustics and directed sampling to identify viable subtidal substrates

Defining and delineating species distribution and habitat is critical to informed management and conservation. This process is complicated in marine environments, where detection of marine taxa and characterization of marine habitat is more difficult. Small pelagic fishes and forage fishes are particularly challenging, though insights may be more accessible in species highly dependent on particular habitat. Pacific sand lance (Ammodytes personatus) is a common and ecologically-important pelagic fish that relies on specific benthic sediments for rest and refuge from predation. We applied multibeam echosounder (MBES) bathymetric data to develop high-definition benthic habitat maps and implemented multiyear sampling to assess potential habitat for sand lance via in situ sampling of sediments. We also applied acoustic Doppler current profilers (ADCP) data and tidally-driven volume-based ocean models to measure current strength and to visualize currents. We leveraged this data to further define and describe habitat for this important forage species. Sediment transport processes were mapped and areas of dispersal, embedment, and accumulation were evaluated. Dynamic bedform habitats, banner banks and glacial banks were identified as potential habitat and sampled for fish presence, density and sediment composition. In the central Salish Sea, approximately 25% of benthic substrates represent potential sand lance habitat. Sand lance prevalence and density correlated with substrate type and sediment coarseness. Densities were highest in areas of coarse grain sediments and presence was limited by fine particulates, such as silt and mud. Tidal currents appear important. Presence and densities of sand lance were correlated with current velocity and distance from current flow path. Nearly all viable sites were located on the immediate margins of high flow (<0.16 km from tidal currents with max speed of 1.72–2.58 m/s). While both flood and ebb were important, processes related to flood were dominant. Viable habitat was not constrained by depth. These results inform a developing atlas for sand lance in the central Salish Sea, provide new insights to subtidal sand lance habitat, characterize conditions that create and maintain subtidal benthic habitat, and provide a template for mapping habitat for this species in the coastal Pacific Ocean.

Enhancing Marine Topography Mapping: A Geometrically Optimized Algorithm for Multibeam Echosounder Survey Efficiency and Accuracy

This research introduces a new multibeam survey line model that optimizes geometric relationships to improve the efficiency and accuracy of surveys over complex seabed topographies. Since existing multibeam echosounder systems have limitations in handling complex terrains, this study presents an advanced model to enhance data quality and operational efficiency. By strategically designing survey lines and optimizing coverage strategies, this paper achieves the optimal configuration of survey lines for secondary measurements in marine areas, ensuring high precision and reliability of measurement data. Experimental results show that the new model significantly outperforms existing technologies in terms of effective coverage area and measurement accuracy, with an average coverage rate of over 95%, higher than existing models, and a weighted average overlap rate of 3.18%, greatly improving the economic efficiency of measurements by reducing redundant coverage and minimizing operational costs. These findings confirm the advantages of the new model in practical applications and offer valuable technical support for advancing seabed mapping technology.

Robust algorithm for automatic surface-based outlier detection in MBES point clouds

Bathymetric multibeam echosounder systems (MBES) provide high-resolution mapping of underwater topography but are highly susceptible to errors due to harsh environmental conditions and the measurement process. Traditionally, manual post-processing is required to ensure data quality, a time-consuming, expensive, and subjective task. To address this issue, we propose a surface-based algorithm for pre-processing and cleaning MBES data that reduces manual intervention and improves consistency. A surface-based algorithm models the underwater topography as a surface instead of processing individual points. By assuming a continuous surface for underwater geometry, the algorithm easily identifies observations that deviate significantly from this model. The method combines a hierarchical B-spline surface with iterative robust estimation to automate data cleaning. Preliminary results on example datasets show a balanced outlier detection accuracy of 0.99, with manual processing time reduced from 2 days to just 30 min.

3D UHR seismic and back-scattering analysis for seabed and ultra-shallow subsurface classification

AbstractRecently, the seabed classification method based on back-scattering data of multi-beam echo-sounder (MBES) is widely used to analyze the distribution of seabed sediment. Although various analysis methods for seabed classification using multi-spectral MBES have been developed, they are limited in securing penetration depth to consider the characteristics of the shallow subsurface structure. In this study, the seabed and ultra-shallow subsurface classification was performed by comparative analysis of box corer sampling, back-scattering, and 2D/3D ultra-high-resolution (UHR) seismic data obtained from Yeongil Bay, South Korea. We proposed a process for seismic ultra-shallow subsurface classification by the segmentation of the primary seabed reflection wavelet and the amplitude analysis. The seabed-reflected amplitude and back-scattering intensity showed similar mapping trends in the relatively homogeneous and thick surface sediment. On the other hand, it was confirmed that back-scattering data and seabed-reflected amplitude show different patterns when the subsurface structure is related to the seabed surface. It is presumed that because seismic data containing relatively low-frequency components have a deeper penetration depth than MBES, they contain more characteristics of the ultra-shallow subsurface than back-scattering data. These were determined that back-scattering has advantages in representing acoustic anomaly distribution by surface sediment type, and seabed-reflected amplitude is advantageous for representing sediment type by ultra-shallow subsurface. In particular, these results were well shown when the surface sediment thinly covered the rocky bottom. Therefore, it is necessary not only to analyze the back-scattering of MBES but also the ultra-shallow subsurface features through seismic data for valid seabed classification.

Source and propagation modelling scenarios for environmental impact assessment: Model verificationa).

Evaluation of possible effects of underwater sound on aquatic life requires quantification of the sound field. A marine sound source and propagation modelling workshop took place in June 2022, whose objectives were to facilitate the evaluation of source and propagation models and to identify relevant metrics for environmental impact assessment. The scope of the workshop included model verification (model-model comparison) and model validation (model-measurement comparison) for multiple sources, including airguns, a low-frequency multi-beam echo sounder, and a surface vessel. Several verification scenarios were specified for the workshop; these are described herein.

Mussel culture monitoring with semi-supervised machine learning on multibeam echosounder data using label spreading

High diversity seabed habitats, such as shellfish aggregations, play a significant role in marine ecosystem sustainability but are susceptible to bottom disturbance induced by anthropogenic activities. Regular monitoring of these habitats with effective mapping methods is therefore essential. Multibeam echosounder (MBES) has been widely used in recent decades for seabed characterization due to its non-destructive manner and extensive spatial coverage compared to traditional methods like bottom sampling. Nevertheless, bottom sampling remains essential to link ground truth with acoustic seabed classification. Using seabed samples and MBES measurements, machine learning techniques are commonly employed to model their relationships and generate classification maps of an extended seabed. However, limited ground truth data, resulting from constraints in regulations, budget, or time, may impede the development of robust machine learning models. To address this challenge, we applied a semi-supervised machine learning method to classify seabed sediments of a blue mussel (Mytilus edulis) cultivation area in the Oosterschelde, the Netherlands. We utilized nine boxcore samples to generate pseudo-labels on MBES data. These pseudo-labels enlarged the training data size, facilitated the training of three comprehensive machine learning algorithms (Gradient Boosting, Random Forest, and Support Vector Machine), and helped to classify the study site into mussel and non-mussel areas. We found the geomorphological and backscatter-related features to be complementary for mussel culture detection. Our classification results were demonstrated effective through expert knowledge of this cultivation area and brought insights for future research on natural mussel habitats.

An Interpretable Multi-Model Machine Learning Approach for Spatial Mapping of Deep-Sea Polymetallic Nodule Occurrences

High-resolution mapping of deep-sea polymetallic nodules is needed (a) to understand the reasons behind their patchy distribution, (b) to associate nodule coverage with benthic fauna occurrences, and (c) to enable an accurate resource estimation and mining path planning. This study used an autonomous underwater vehicle to map 37 km2 of a geomorphologically complex site in the Eastern Clarion–Clipperton Fracture Zone. A multibeam echosounder system (MBES) at 400 kHz and a side scan sonar at 230 kHz were used to investigate the nodule backscatter response. More than 30,000 seafloor images were analyzed to obtain the nodule coverage and train five machine learning (ML) algorithms: generalized linear models, generalized additive models, support vector machines, random forests (RFs) and neural networks (NNs). All models ML yielded similar maps of nodule coverage with differences occurring in the range of predicted values, particularly at parts with irregular topography. RFs had the best fit and NNs had the worst spatial transferability. Attention was given to the interpretability of model outputs using variable importance ranking across all models, partial dependence plots and domain knowledge. The nodule coverage is higher on relatively flat seafloor ( < 3°) with eastward-facing slopes. The most important predictor was the MBES backscatter, particularly from incident angles between 25 and 55°. Bathymetry, slope, and slope orientation were important geomorphological predictors. For the first time, at a water depth of 4500 m, orthophoto-mosaics and image-derived digital elevation models with 2-mm and 5-mm spatial resolutions supported the geomorphological analysis, interpretation of polymetallic nodules occurrences, and backscatter response.