Hydrophone Data Research Articles

Gridless DOA Estimation Method for Arbitrary Array Geometries Based on Complex-Valued Deep Neural Networks

Gridless direction of arrival (DOA) estimation methods have garnered significant attention due to their ability to avoid grid mismatch errors, which can adversely affect the performance of high-resolution DOA estimation algorithms. However, most existing gridless methods are primarily restricted to applications involving uniform linear arrays or sparse linear arrays. In this paper, we derive the relationship between the element-domain covariance matrix and the angular-domain covariance matrix for arbitrary array geometries by expanding the steering vector using a Fourier series. Then, a deep neural network is designed to reconstruct the angular-domain covariance matrix from the sample covariance matrix and the gridless DOA estimation can be obtained by Root-MUSIC. Simulation results on arbitrary array geometries demonstrate that the proposed method outperforms existing methods like MUSIC, SPICE, and SBL in terms of resolution probability and DOA estimation accuracy, especially when the angular separation between targets is small. Additionally, the proposed method does not require any hyperparameter tuning, is robust to varying snapshot numbers, and has a lower computational complexity. Finally, real hydrophone data from the SWellEx-96 ocean experiment validates the effectiveness of the proposed method in practical underwater acoustic environments.

PickBlue: Seismic Phase Picking for Ocean Bottom Seismometers With Deep Learning

AbstractDetecting phase arrivals and pinpointing the arrival times of seismic phases in seismograms is crucial for many seismological analysis workflows. For land station data, machine learning methods have already found widespread adoption. However, deep learning approaches are not yet commonly applied to ocean bottom data due to a lack of appropriate training data and models. Here, we compiled an extensive and labeled ocean bottom seismometer (OBS) data set from 15 deployments in different tectonic settings, comprising ∼90,000 P and ∼63,000 S manual picks from 13,190 events and 355 stations. We propose PickBlue, an adaptation of the two popular deep learning networks EQTransformer and PhaseNet. PickBlue joint processes three seismometer recordings in conjunction with a hydrophone component and is trained with the waveforms in the new database. The performance is enhanced by employing transfer learning, where initial weights are derived from models trained with land earthquake data. PickBlue significantly outperforms neural networks trained with land stations and models trained without hydrophone data. The model achieves a mean absolute deviation of 0.05 s for P‐waves and 0.12 s for S‐waves, and we apply the picker on the Hikurangi Ocean Bottom Tremor and Slow Slip OBS deployment offshore New Zealand. We integrate our data set and trained models into SeisBench to enable an easy and direct application in future deployments.

Methodology of elastic full‐waveform inversion of multicomponent ocean‐bottom data for anisotropic media

AbstractFull‐waveform inversion of multicomponent data can provide an improved estimation of medium parameters using both compressional‐ and shear‐wave information. However, most earlier studies that involved a full‐waveform inversion of ocean‐bottom data are based on acoustic anisotropic or elastic isotropic approximations. Here, we consider realistic elastic anisotropic media and develop an efficient full‐waveform inversion framework for estimating model parameters. We simulate seismic wavefields using a previously developed coupled acoustic/elastic wave propagator that implements a mimetic finite‐difference method with fully staggered grids to accurately handle the fluid/solid boundary conditions. The algorithm employs a multiscale approach starting from low frequencies and incorporating higher frequency bands in the later inversion stages. We analyse the influence of different types of input data on the accuracy of the inverted anisotropy parameters for hard and soft water bottoms. The employed misfit function incorporates information from both hydrophones and ocean‐bottom geophones. Numerical examples indicate that injecting multiple data components simultaneously increases the complexity of the objective function and often degrades the quality of the estimated medium parameters. Thus, we propose a sequential strategy using a single data component at a time. Pressure (hydrophone) data alone can provide satisfactory results if long offsets (i.e., with the offset/depth ratio ≥ 3) are available. Adding the horizontal particle‐displacement or ‐velocity components increases the accuracy of the estimated shear‐wave vertical velocity () and P‐wave normal‐moveout () velocity, especially for strongly heterogeneous sub‐water‐bottom models.

Performance of distributed acoustic sensing relative to co-located hydrophone measurements

Distributed acoustic sensing (DAS) is a technology that transforms telecommunication fiber optic cables into dense sensor arrays by continuously transmitting pulses of light down the cable and measuring backscattering from natural inhomogeneities within the fiber cable. The technology can densely sample the acoustic field over long ranges (up to 100 km), providing a means for large scale passive acoustic monitoring. The DAS Calibration Experiment 2022 (DASCAL22) provides 9 days of co-located hydrophone and DAS data in Puget Sound, WA. The DAS data were recorded with a 2kHz sampling rate over ∼3.5 km of cable in a channel reaching close to 100 m depth and with DAS channels spaced 6.38 m. A combination of an active source broadcast from 1, 5, and 10 m depths, and passive signals (primarily boat noise) are used to investigate the capabilities of DAS to sense a range of frequencies and the impacts of the environment on those capabilities.

Arnoux’s beaked whales acoustically misidentified as Type C killer whales in Prydz Bay, Antarctica

The Southern Ocean is home to five ecotypes of killer whale, which differ in diet, morphology, behaviour, and vocal repertoire. Recently, fish-eating Type C killer whales were reported in Prydz Bay, Antarctica from sounds recorded in hydrophone data collected in March 2017, on the basis that these sounds bore similarities to acoustic behaviour of Icelandic fish-eating killer whales. Here we re-examine the structure of the sounds identified as Prydz Bay Type C killer whales and we identify the calls to be those of the Arnoux’s beaked whale. The whistles are more structurally similar and fall within the same frequency range as those of the Arnoux’s beaked whale, rather than the killer whale. Of the two killer whale types observed in Prydz Bay, the most common had behaviour and morphology typical of Type B killer whales, while the other had morphology typical of Type A. Although we show that the sounds are not likely to be killer whale Type C calls, this would be the first recorded occurrence of Arnoux’s beaked whales in Prydz Bay. Arnoux’s beaked whales are a rarely seen, cryptic whale species.

Signal recovery and noise suppression of the Ocean‐Bottom Cable P‐component data based on improved dense convolutional network

AbstractEffective attenuation of noise in seismic data is important for high‐quality seismic imaging. Noise suppression in Ocean‐Bottom Cable data is particularly challenging. The challenge for the geophysicist is to process the individual hydrophone and vertical geophone data up to a level where they can conveniently be combined for effective multiple suppressions. In this study, we propose a deep learning‐based solution for noise attenuation and signal recovery of the P‐component of Ocean‐Bottom Cable data. To effectively attenuate complex noise, a denoising model based on dense convolutional network is proposed for Ocean‐Bottom Cable data processing. The backbone of the denoising network uses dense blocks to extract the potential features. Dense connections are applied to fuse the features at each stage to further enhance the effective information and thus improve the reconstruction of the signal. A high‐quality training set was built for the training network to ensure that the trained model was suitable for noise suppression. Synthesis and field experiments show that the proposed method can completely eliminate complex noise and recover weak signals from the P‐component data of the Ocean‐Bottom Cable data.

Surface and Underwater Acoustic Source Recognition Using Multi-Channel Joint Detection Method Based on Machine Learning

Sound source recognition is a very important application of passive sonar. How to distinguish between surface and underwater acoustic sources has always been a challenge. Due to the mixing of underwater target radiated noise and marine environmental noise, especially in shallow water environments where multipath effects exist, it is difficult to distinguish them. To solve the surface and underwater acoustic source recognition problem, this paper proposes a multi-channel joint detection method based on machine learning. First, the simulation data are generated using the normal model KRAKEN setting in the same environment as the SACLANT 1993 experiment, which uses a vertical linear array of 48 hydrophones. Secondly, the GBDT classifier and LightGBM classifier are trained separately, and then the model is evaluated using precision, recall, F1, and accuracy. Finally, four ML models (kNN, random subspace kNN, GBDT, and LightGBM) are used to analyze all 48 channels of hydrophone data. For each model, two kinds of feature extraction methods (module features, real and imaginary features) are applied. Generally, the results show that both GBDT and LightGBM models have better performance than both kNN and random subspace kNN ones. For both GBDT and LightGBM models, the results using module features have better performance than using real and imaginary features.

Distributed acoustic sensing for detecting near surface hydroacoustic signals.

Distributed acoustic sensing (DAS) is a technology that turns a fiber-optic cable into an acoustic sensor by measuring the phase change of backscattered light caused by changes in strain from an acoustic field. In October 2022, 9 days of DAS and co-located hydrophone data were collected in the Puget Sound near Seattle, WA. Passive data were continuously recorded for the duration and a broadband source was fired from several locations and depths on the first and last days. This dataset provides comparisons between DAS and hydrophone measurements and demonstrates the ability of DAS to measure acoustics signals up to ∼700 Hz.

Geographic and seasonal variation of blue whale calls in the North Pacific

The call characteristics and distribution of blue whales in the North Pacific were examined by use of acoustic surveys. Two distinct vocalisation types have been previously attributed to blue whales from limited regions in the North Pacific (cf. Thompson and Friedl, 1982; Rivers, 1997). Hydrophone data from sixteen sites in the North Pacific were examined for these blue whale vocalisations. There were distinct geographic and seasonal differences between the occurrence of the two vocalisation types. The hydrophones that were more westerly recorded the ‘northwestern’ Pacific vocalisation, those in the eastern Pacific recorded the ‘northeastern’ Pacific vocalisation and those in the central Pacific recorded both types. Northeastern vocalisations were recorded from July-December in the northeast Pacific and February-May in the eastern tropical Pacific. Northwestern vocalisations were recorded most often from July-December, and were essentially absent from March-May in the northwestern Pacific. These results suggest that the different vocalisation types may represent at least two distinct groups of blue whales in the North Pacific.

Active acoustic surveys reveal coastal fish community resistance to an environmental perturbation in South Florida.

Coastal fish communities are under increasing levels of stress associated with climate variation and anthropogenic activities. However, the high degree of behavioral plasticity of many species within these communities allow them to cope with altered environmental conditions to some extent. Here, we combine meteorological information, data from hydroacoustic surveys, and recordings of goliath grouper sound production to examine the response of coastal fish communities to heavy rainfall events in South Florida, USA, that resulted in the release of excess storm water into surrounding estuaries and coastal waters. We observed a nearly 12,000% increase in water column acoustic backscatter following a heavy rainfall event of September 16th, 2015. Interestingly, estimates of school backscatter, a proxy for biomass, increased by 172% with the onset of the perturbation. Schooling fish density also increased by 182%, as did acoustically derived estimates of mean schooling fish length (21%). Following the perturbed period, school backscatter decreased by 406%, along with schooling density (272%), and mean schooling fish length (35%). Hydrophone and hydroacoustic data also revealed that goliath grouper (Epinephelus itajara) spawning aggregations were persistent in the region throughout the duration of the study and continued to exhibit courtship behavior during the perturbed period. Our observations demonstrate the high level of resistance common in coastal species but raises new questions regarding the threshold at which fish communities and reproductive activities are disrupted. As coastal land use continues to increase, and the effects of global climate change become more pronounced, more Before-After Control Impact (BACI) studies will provide improved insight into the overall response of nearshore communities to future perturbations and the cumulative effect of repeated perturbations over extended periods.

A processing for ocean-bottom multicomponent data with seismic interferometry: a case study of southern offshore in Korea

Marine seismic surveys are widely used to explore oil and gas resources and monitor CO2 storage. In particular, ocean-bottom seismic surveys allow the acquisition of high-quality data without the influence of the water-layer environment. However, ocean-bottom seismic surveys are more expensive than streamer surveys and P-wave survey data exhibit free-surface multiples, which are persistent problems. PZ summation can be applied to process multicomponent ocean-bottom seismic data, which effectively reduces ghosting and water-layer reverberation by combining hydrophone and vertical geophone data. In this study, we processed and analysed ocean-bottom cable data from a shallow gas area offshore from Korea. In our field-survey simulation, PZ summation was applied to the synthetic data to separate up-going and down-going wavefields and attenuate free-surface multiples. In addition, improved image can be obtained compared to conventional stack section by applying seismic interferometry to the acquired up-going wavefield data.

Seismic airgun sound propagation in shallow water of the East Siberian shelf and its prediction with the measured source signature

Seismic airgun sound was measured with an autonomous passive acoustic recorder as a function of distance from 18.6 to 164.2 km in shallow water (<70 m) at the continental shelf of the East Siberian Sea in September 2019. The least-square regression curves were derived in the zero-to-peak sound pressure level, sound exposure level, and band level in a frequency range between 10 and 300 Hz using the initial amplitude scaled from the near-field hydrophone data. In addition, propagation modeling based on the parabolic equation with the measured source spectrum was performed for range-dependent bathymetry, and the results were compared with the band level of the measurements. The sediment structure of the measurement area was a thin layer of iceberg-scoured postglacial mud overlying a fast bottom with high density based on grounding events of past ice masses. The observed precursor arrivals, modal dispersion, and rapid decrease in spectrum level at low frequencies can be explained by the condition of the high-velocity sediment. Our results can be applied to studies on the inversion of ocean boundary conditions and measurement geometry and basic data for noise impact assessment.

Hydrovolcanic Explosions at the Lava Ocean Entry of the 2018 Kilauea Eruption Recorded by Ocean-Bottom Seismometers

AbstractFrom the beginning of May 2018, the Kilauea Volcano on the island of Hawaii experienced its largest eruption in 200 yr followed by a period of unrest for months. Because hot molten lava entered the ocean from the ocean-entry point near the lower East Rift Zone, the lava–water interaction led to explosions. Some explosions were near the water surface and ejected fragments of lava, also known as lava bombs. In the early morning on 16 July 2018, one of those lava bombs, which was almost the size of a basketball, hit a sightseeing boat and injured 23 people. In this study, we analyzed the hydrophone data recorded from July to mid-September by ocean-bottom seismometers (OBSs) deployed offshore near the ocean entry point to identify and locate the hydroacoustic signals of the lava–water explosions. Acoustic signals of hydrovolcanic explosions are characterized by a short duration (less than a few seconds) and a broad frequency range (at least up to 100 Hz). To automate event detection, a short-term average versus long-term average method was applied to the complete dataset. Approximately 4300 events were detected and located near the coastline and further used to prepare a catalog. The distribution of the lava–water explosions is consistent with the pattern of the offshore lava delta formed during the 2018 eruption. Identifying such hydroacoustic signals recorded by OBSs may provide new avenues of research using various seismoacoustic events associated with volcanic eruptions.

Can elastic waveform inversion benefit from inverting multicomponent data?

Waveform inversion aims to retrieve high-resolution earth parameter volumes. Due to its high computational cost, acoustic approaches using low to mid seismic data frequencies are often applied in velocity model building. However, in the presence of large parameter contrasts, an elastic formulation should be preferred due to wave interferences that limit the applicability of phase- and kinematics-only approaches. The benefits of elastic waveform inversion have been demonstrated with inversion of hydrophone data in marine environments. Here, we extend the approach to inversion of the vertical geophone component. We propose a weighted cost function in the waveform inversion algorithm to jointly invert multicomponent data sets, and we compare the results to the inversion of single-component data. For this study, we use an ocean-bottom-node data set from the deepwater Gulf of Mexico around a salt dome. We show that slightly different velocity models, reverse time migration images, and angle gathers are retrieved when using hydrophone or vertical geophone data, further improving either the shallow or deeper sediments. The joint inversion combines the improvements brought by the single-component inversions. Though it doubles the cost, joint elastic waveform inversion of hydrophone and vertical geophone data can help velocity model building around salt bodies.

Comparing distributed acoustic sensing data with hydrophone recordings

Distributed acoustic sensing (DAS) is a technology that transforms telecommunication fiber optic cables into dense sensor arrays by continuously transmitting pulses of light down the cable and measuring backscattering from natural inhomogeneities within the fiber cable. The technology can densely sample the acoustic field over long ranges (up to 100 km), providing a means for large scale passive acoustic monitoring. To evaluate the capabilities of DAS, it is necessary to benchmark and calibrate the technology relative to traditional hydrophone data. The DAS Calibration Experiment 2022 (DASCAL22) recorded 9 days of both DAS and hydrophone data in Puget Sound, WA in October 2022. The DAS data were recorded with a sample rate of 2 kHz, and the cable extended 3.5 km on the seafloor between two islands, reaching depths of 100 m, and the hydrophones were moored adjacent to the DAS cable at 5 m and 25 m from the seafloor. The recordings include impulses from an active source at 1 m, 5 m, and 10 m depths, and an abundance of passive acoustic data corresponding to ship traffic, wind, and rain. This work aims to draw comparisons between the hydrophone and DAS recordings to evaluate the capability of DAS at detecting sounds at frequencies as high as 1 kHz.

Is 3D frequency-domain FWI of full-azimuth/long-offset OBN data feasible? The Gorgon data FWI case study

Frequency-domain full-waveform inversion (FWI) is potentially amenable to efficient processing of full-azimuth long-offset stationary-recording seabed acquisition carried out with a sparse layout of ocean-bottom nodes (OBNs) and broadband sources because the inversion can be performed with a few discrete frequencies. However, computing the solution of the forward (boundary-value) problem efficiently in the frequency domain with linear algebra solvers remains a challenge for large computational domains involving tens to hundreds of millions of parameters. We illustrate the feasibility of 3D frequency-domain FWI with a subset of the 2015/2016 Gorgon OBN data set in the North West Shelf, Australia. We solve the forward problem with the massively parallel multifrontal direct solver MUMPS, which includes four key features to reach high computational efficiency: an efficient parallelism combining message-passing interface and multithreading, block low-rank compression, mixed-precision arithmetic, and efficient processing of sparse sources. The Gorgon subdata set involves 650 OBNs that are processed as reciprocal sources and 400,000 sources. Monoparameter FWI for vertical wavespeed is performed in the viscoacoustic vertically transverse isotropic approximation with a classical frequency continuation approach proceeding from a starting frequency of 1.7 Hz to a final frequency of 13 Hz. The target covers an area ranging from 260 km2 (frequency ≥ 8.5 Hz) to 705 km2 (frequency ≤ 8.5 Hz) for a maximum depth of 8 km. Compared to the starting model, FWI dramatically improves the reconstruction of the bounding faults of the Gorgon horst at reservoir depths as well as several intrahorst faults and several horizons of the Mungaroo Formation down to a depth of 7 km. Seismic modeling reveals a good kinematic agreement between recorded and simulated data, but amplitude mismatches between the recorded and simulated reflection from the reservoir suggest elastic effects. Therefore, future works involve multiparameter reconstruction for density and attenuation before considering elastic FWI from hydrophone and geophone data.

Active and Passive Near-Field Hydrophones Used in Ultrashallow Waters

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 207728, “Using Active and Passive Near-Field Hydrophones To Image the Near-Surface in Ultrashallow Waters Offshore Abu Dhabi,” by Oleg Khakimov, Yaser Gholami, and Bertrand Tertrais, CGG, et al. The paper has not been peer reviewed. _ Seismic surveys can leave the near-surface undersampled. This is particularly a challenge offshore Abu Dhabi, where a complex near-surface contaminates the seismic image. To mitigate these effects, the authors use near-field hydrophone (NFH) data. NFHs are positioned approximately a meter above each airgun and are designed to record near-field pressure. The authors propose combining active and passive NFH measurements, using active NFHs for the very shallow section and passive NFHs for the deeper section. The authors have applied this technique to a recent node survey acquired offshore Abu Dhabi. Introduction Ocean-bottom-node (OBN) acquisition in shallow-water environments has many benefits. One technique involves imaging using NFH data. The main purpose of an NFH is to monitor proper functioning of the aiguns during acquisition. They also are used often to compute the far-field signature using the concept of notional sources. But the NFHs also record reflections from the near surface and can be used to construct an image. Two types of NFH are recorded: passive, when the recording hydrophones are on top of the array that is currently not firing; and active, when the hydrophones are on top of the firing arrays. Usually, passive NFHs are the preferred choice for imaging because of less-energetic direct arrival contamination; however, in extremely shallow environments, the distance between sources and receivers for the passive NFHs is too large to capture reflections from the water bottom and reflectors immediately below. In such situations, active NFHs have the advantage of recording zero-offset data. The sampling rate of the NFHs in this study was 1 ms, which allowed imaging of frequencies above 250 Hz. Careful processing is required for active NFH data because it is strongly contaminated by the near-field signature. To extend the maximum depth of the image, active NFHs were combined with passive. The complete paper presents work performed on NFH data acquired during OBN acquisition in Abu Dhabi shallow waters. The data examples come from a priority area where many tests were performed to select an optimal work flow. The total size of the study area was approximately 790 km2. Acquisition Parameters Data were acquired using five vessels. OBN acquisition involved blended acquisition, with four vessels operating at the same time with a minimum distance separation of 6 km. Vessels moved at an average speed of 9 km/h, which provided approximately 5-second time separation between shots. Two of the vessels had dual-source arrays, and the other two had only single arrays. In the shallower areas, only smaller vessels with single arrays were shooting. This meant that passive NFH data acquisition was not applicable everywhere.

The role of sea ice in the distribution, habitat use, and phenology of eastern North Pacific gray whales

A strong negative relationship has been observed between calf production of eastern North Pacific gray whales Eschrichtius robustus and Pacific Arctic sea ice area with the exception of 2 potentially anomalous periods: 2013-2014 and 2017-2019. Sea ice may play a role in reproductive variability by blocking gray whale access to foraging hotspot habitats. We tested this ‘sea ice exclusion’ hypothesis using complementary aerial survey and moored hydrophone data sets that together provide broad spatial and temporal coverage. A negative relationship (generalized additive model, p < 0.001) was found between gray whale aerial counts and ice concentration in the northeast Chukchi Sea (2008-2019), with a nonlinear increase in negative slope above 45-55% concentration. A comparison of distribution patterns between years further indicated the absence of sightings within northern Bering and northeastern Chukchi foraging hotspots during years with delayed ice break-up versus years with an early-to-average ice retreat. Passive acoustic recordings likewise showed the near-absence of detections during periods of dense ice cover and indicated a strongly positive relationship between ice break-up dates and the timing of acoustic detection onset. Together, these analyses point to the important role of sea ice in gray whale distribution, habitat use, and phenology. However, a relatively consistent 10-15 d lag observed between ice break-up and acoustic onset dates suggested that the mechanism(s) underpinning gray whale interactions with sea ice may be more complex than simply ‘sea ice exclusion.’ Here, we propose ‘prey quality timing’ as a potential alternative hypothesis that warrants further investigation.

Supervised Learning Approach for Multi-Component Deghosting with a Focus on Shallow Water Data

In recent years, the use of machine learning (ML) has shown more and more successful applications, initially mainly in seismic interpretation (classification) but recently also in seismic processing challenges (regression) like denoising, interpolation and demultiple. In this paper, a supervised machine learning algorithm is proposed for receiver deghosting on multi-component data. A neural network is trained to map input pairs of hydrophone and geophone data into the deghosted up-going wavefield. This data-driven approach could offer a user-friendly alternative to existing deghosting methods, like the PZ-summation approach that is commonly used for multi-component streamer and ocean-bottom acquisitions. Although the PZ-summation approach offers an analytical solution to the deghosting problem, it requires some pre-processing steps that make it more time-consuming and user-intensive to run and QC. The advantage of the proposed ML approach is that it can be directly applied to multi-component data without any parameterization. However, the success of the supervised method strongly depends on whether a neural network can be trained easily with minimal user-interaction and can be applied to various types of (field) data without adapting the network for each individual case. To investigate if this is possible for the supervised multi-component receiver deghosting method, the selected neural network is trained on easy to model 1.5D data examples and applied to test data with increasing complexity. Results are very encouraging and evaluated to be at least on par in comparison to the existing PZ-summation deghosting method.

Source and receiver deghosting by demigration‐based supervised learning

AbstractDeghosting of marine seismic data is an important and challenging step in the seismic processing flow. We describe a novel approach to train a supervised convolutional neural network to perform joint source and receiver deghosting of single‐component (hydrophone) data. The training dataset is generated by demigration of stacked depth migrated images into shot gathers with and without ghosts using the actual source and receiver locations from a real survey. To create demigrated data with ghosts, we need an estimate of the depth of the sources and receivers and the reflectivity of the sea surface. In the training process, we systematically perturbed these parameters to create variability in the ghost timing and amplitude and show that this makes the convolutional neural network more robust to variability in source/receiver depth, swells and sea surface reflectivity. We tested the new method on the Marmousi synthetic data and real North Sea field data and show that, in some respects, it performs better than a standard deterministic deghosting method based on least‐squares inversion in the τ‐p domain. On the synthetic data, we also demonstrate the robustness of the new method to variations in swells and sea‐surface reflectivity.