Hugo Latorre
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Hugo Latorre has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Sergi Ventosa
Institute of Marine Sciences, CSIC, Barcelona, Spain
4 shared publications Arantza Ugalde
3 shared publications- Arantza Ugalde
Institute of Marine Sciences, ICM-CSIC, Barcelona, Spain
2 shared publications - Melania Cubas Armas
Institute of Marine Sciences, CSIC, Barcelona, Spain
2 shared publications - Aleix Seguí
ETH Zurich, Zurich, Switzerland
1 shared publication - Antonio Villaseñor
Institute of Marine Sciences, ICM-CSIC, Barcelona, Spain
1 shared publication
Latest Publications
- Open Access
- 10 June 2026
- Published by Copernicus GmbH
Adapting a multiscale phase picking algorithm to detect whale songs in marine environments
While sensing marine environments, seismic and DAS instruments routinelyrecord hydroacoustic signals together with transient seismic phases. Although someof these signals correspond to converted phases at the seafloor boundary, othersignals that originate within the water column are also recorded. Whale vocalisationsare a particular class of these hydroacoustic signals, commonly referred to as whalesongs in marine science.In the case of fin whales, distinct vocalisation types include 20-Hz andbackbeat reproductive calls produced by males, and 40-Hz feeding-associated callsattributed to both sexes. All of them fall within the bandwidth and samplingcharacteristics commonly available in seismic and DAS experiments and aretherefore relevant for bioacoustics monitoring. Unlike seismic arrivals correspondingto P and S phases, which are typically short and impulsive, individual notes sung bywhales are composed of many cycles. Although existing picking algorithms canalready detect some of these notes, often from amplitude increases, detectionperformance can be improved by developing strategies that account for the narrow-band nature and longer duration of these signals.Here we adapt the Kurtosis-Value-Picker (KVP) algorithm, originallydeveloped by the authors to pick P and S phases with accurate arrival times, to betterdetect individual notes within whale songs. Since accurate picking times are not ascritical as detection itself for these particular types of signals and their later study, wecan instead focus on their specific frequency content and waveform. Replacing theRicker wavelet used by KVP with the Morlet wavelet, we find that detectionimproves significantly in tested data and that non-target signals are more effectivelyrejected. The time-frequency resolution trade-off introduced by the Morlet wavelet isnot limiting when the focus is on detection rather than accurate picking times. Thisallows for better narrow-band selection, which in turn facilitates improvedclassification of notes within whale songs.
- 10 June 2026
- Published by Copernicus GmbH
Open AccessAdapting a multiscale phase picking algorithm to detect whale songs in marine environments
While sensing marine environments, seismic and DAS instruments routinelyrecord hydroacoustic signals together with transient seismic phases. Although someof these signals correspond to converted phases at the seafloor boundary, othersignals that originate within the water column are also recorded. Whale vocalisationsare a particular class of these hydroacoustic signals, commonly referred to as whalesongs in marine science.In the case of fin whales, distinct vocalisation types include 20-Hz andbackbeat reproductive calls produced by males, and 40-Hz feeding-associated callsattributed to both sexes. All of them fall within the bandwidth and samplingcharacteristics commonly available in seismic and DAS experiments and aretherefore relevant for bioacoustics monitoring. Unlike seismic arrivals correspondingto P and S phases, which are typically short and impulsive, individual notes sung bywhales are composed of many cycles. Although existing picking algorithms canalready detect some of these notes, often from amplitude increases, detectionperformance can be improved by developing strategies that account for the narrow-band nature and longer duration of these signals.Here we adapt the Kurtosis-Value-Picker (KVP) algorithm, originallydeveloped by the authors to pick P and S phases with accurate arrival times, to betterdetect individual notes within whale songs. Since accurate picking times are not ascritical as detection itself for these particular types of signals and their later study, wecan instead focus on their specific frequency content and waveform. Replacing theRicker wavelet used by KVP with the Morlet wavelet, we find that detectionimproves significantly in tested data and that non-target signals are more effectivelyrejected. The time-frequency resolution trade-off introduced by the Morlet wavelet isnot limiting when the focus is on detection rather than accurate picking times. Thisallows for better narrow-band selection, which in turn facilitates improvedclassification of notes within whale songs.
- Open Access
- 13 March 2026
- Published by Copernicus GmbH
Spectral analysis of background and transient signals at Mount Etna using rectilinear fibre-optic segments
Etna is the largest, most active and closely monitored volcano in Europe,making it a crucial study region for volcanology and geohazard assessment. In earlyJuly 2019, a 1.5 km fibre-optic cable was deployed near the summit of Mount Etnaand interrogated for two months. The cable was divided into four main segments, twoof which point towards different active crater areas. Temporary seismic broadbandstations and infrasound sensors were also deployed along the cable. During theexperiment, three distinct eruptive events were recorded. The first two events arecharacterised by a large number of explosions in the active crater area, together withan increase in background tremor activity. The third event is characterised by a largerincrease in background tremor, but almost no explosions.The continuous recordings are analysed in the frequency-wavenumber domain,which reveals the features of the background tremor activity and the stacked transientsignals, such as explosions. During the first two eruptive events, the stack ofexplosive sources is characterised by a non-dispersive arrival, travelling withdifferent apparent velocities along each segment, and a non-linear ground response upto 25 Hz. These segments can be used as an antenna to estimate an average back-azimuth for the explosions, which come from the same crater area during botheruptive events.Outside of the three eruptive events, the background tremor features two slowdispersion modes, both well resolved on the raw recordings. The slowest mode isaffected by gauge-length attenuation at higher frequencies, due to its shortwavelength, but remains detectable up to 27 Hz, with group velocities as low as 170m/s. These observations showcase the utility of simple, rectilinear geometries indeployments despite their known shortcomings, such as in location procedures. Forknown source regions, such as volcanoes, a well-oriented segment can leveragecontinuous activity to record the incoming wavefield and extract dipersion curveswithout the need to perform cross-correlations, simplifying the workflow.
- 13 March 2026
- Published by Copernicus GmbH
Open AccessSpectral analysis of background and transient signals at Mount Etna using rectilinear fibre-optic segments
Etna is the largest, most active and closely monitored volcano in Europe,making it a crucial study region for volcanology and geohazard assessment. In earlyJuly 2019, a 1.5 km fibre-optic cable was deployed near the summit of Mount Etnaand interrogated for two months. The cable was divided into four main segments, twoof which point towards different active crater areas. Temporary seismic broadbandstations and infrasound sensors were also deployed along the cable. During theexperiment, three distinct eruptive events were recorded. The first two events arecharacterised by a large number of explosions in the active crater area, together withan increase in background tremor activity. The third event is characterised by a largerincrease in background tremor, but almost no explosions.The continuous recordings are analysed in the frequency-wavenumber domain,which reveals the features of the background tremor activity and the stacked transientsignals, such as explosions. During the first two eruptive events, the stack ofexplosive sources is characterised by a non-dispersive arrival, travelling withdifferent apparent velocities along each segment, and a non-linear ground response upto 25 Hz. These segments can be used as an antenna to estimate an average back-azimuth for the explosions, which come from the same crater area during botheruptive events.Outside of the three eruptive events, the background tremor features two slowdispersion modes, both well resolved on the raw recordings. The slowest mode isaffected by gauge-length attenuation at higher frequencies, due to its shortwavelength, but remains detectable up to 27 Hz, with group velocities as low as 170m/s. These observations showcase the utility of simple, rectilinear geometries indeployments despite their known shortcomings, such as in location procedures. Forknown source regions, such as volcanoes, a well-oriented segment can leveragecontinuous activity to record the incoming wavefield and extract dipersion curveswithout the need to perform cross-correlations, simplifying the workflow.
- Open Access
- 28 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 18 (19391404) , 19869-19883
Automatic Identification of High-Quality Channels in Distributed Acoustic Sensing Through Implementation of a Channel Quality Index
Using fiber-optic cables, distributed acoustic sensing (DAS) has emerged as a powerful technology for seismic monitoring, particularly in oceanic regions, but the massive amount of data generated presents challenges for its integration into operational seismic networks. However, since the quality of DAS recordings varies along the fiber due to physical and environmental factors, not all data collected from fiber-optic cables contribute meaningfully to earthquake detection and characterization. Identifying and filtering out low-quality channels reduces unnecessary data processing and optimizes resource use. To address this, we developed a channel quality index (CQI), a machine-learning-based approach that automatically identifies high-quality DAS channels based on their ability to record earthquake signals. Trained on data from four offshore DAS experiments in the Spanish Mediterranean and Canary Islands, our model evaluates temporal, frequency, and energy-based features to assign quality scores to individual channels, reliably identifying high-quality channels with high precision (0.92–0.96) and strong performance in terms of area under the receiver operating characteristic curve (ROC–AUC, 0.92–0.98). By prioritizing high-quality data, CQI enhances the efficiency of data transmission and storage in large-scale DAS deployments. It enables the selective retrieval of relevant seismic data from repositories, reducing storage demands and minimizing data transmission time, which is especially critical for large datasets. It also provides a mechanism to quantify information loss in data compression strategies, ensuring critical seismic signals are preserved. These capabilities of CQI facilitate the practical adoption of DAS for both real-time and long-term seismic monitoring, optimizing data quality and resource management.
- 28 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 18 (19391404) , 19869-19883
Open AccessAutomatic Identification of High-Quality Channels in Distributed Acoustic Sensing Through Implementation of a Channel Quality Index
Using fiber-optic cables, distributed acoustic sensing (DAS) has emerged as a powerful technology for seismic monitoring, particularly in oceanic regions, but the massive amount of data generated presents challenges for its integration into operational seismic networks. However, since the quality of DAS recordings varies along the fiber due to physical and environmental factors, not all data collected from fiber-optic cables contribute meaningfully to earthquake detection and characterization. Identifying and filtering out low-quality channels reduces unnecessary data processing and optimizes resource use. To address this, we developed a channel quality index (CQI), a machine-learning-based approach that automatically identifies high-quality DAS channels based on their ability to record earthquake signals. Trained on data from four offshore DAS experiments in the Spanish Mediterranean and Canary Islands, our model evaluates temporal, frequency, and energy-based features to assign quality scores to individual channels, reliably identifying high-quality channels with high precision (0.92–0.96) and strong performance in terms of area under the receiver operating characteristic curve (ROC–AUC, 0.92–0.98). By prioritizing high-quality data, CQI enhances the efficiency of data transmission and storage in large-scale DAS deployments. It enables the selective retrieval of relevant seismic data from repositories, reducing storage demands and minimizing data transmission time, which is especially critical for large datasets. It also provides a mechanism to quantify information loss in data compression strategies, ensuring critical seismic signals are preserved. These capabilities of CQI facilitate the practical adoption of DAS for both real-time and long-term seismic monitoring, optimizing data quality and resource management.
- Open Access
- 15 May 2025
- Published by Copernicus GmbH
A Pragmatic Approach for Integrating Submarine DAS Data with Onshore Seismometer Networks: A Case Study in the Canary Islands
Distributed Acoustic Sensing (DAS) technology offers a valuable opportunity to enhance seismological monitoring of ocean floors. Typically, in a standard monitoring network, most seismic stations are installed on land, with only a few ocean-bottom seismometers available. Consequently, monitoring oceanic seismicity is intrinsically challenging due to the lack of observations near seismic sources. In this context, DAS systems, for instance, those installed on fibers connecting islands and land-islands, can help bridge this observational gap. However, integrating the spatially dense DAS data (meter scale) with the often sparser seismometer data (kilometer scale) is not straightforward. Specifically, inverting DAS alongside seismometer P- and S-wave arrival times can lead to biased location results due to the numerical disparity between datasets and/or outliers not identified. Moreover, employing the full set of DAS arrival times can be computationally intensive, limiting its feasibility for real-time monitoring and integration into routine seismological software.Automated weighting methods can help mitigate bias introduced by arrival time outliers in data inversion. This is particularly useful for DAS data, where user control over individual channels is limited. However, suppose the goal is to use DAS as a complementary tool to a seismometer network, and meter-scale spatial density is not essential. In that case, DAS data selection/sub-sampling can improve computational efficiency. To this end, we propose an approach that, for a given seismological network and DAS system, a) automatically identifies “reliable” DAS channels using a machine learning classifier trained on specific data attributes and b) further selects a subset of DAS channels to achieve similar interchannel spacing to the network. The proposed strategy generates a final set of DAS P- and S-wave arrival times with a number of observations comparable to the network. To test the benefits of this procedure on oceanic seismic monitoring, we use data from a fiber optic cable northeast of Gran Canaria and seismometers operated by the Instituto Geográfico Nacional (IGN) in the Canary Islands. We then compare event locations obtained using: a) IGN-only P- and S-wave arrival times, b) DAS-only P- and S-wave arrival times (unselected), and c) IGN and selected DAS P- and S-wave arrival times using the proposed method. Event locations are estimated using a hierarchical Markov Chain Monte Carlo approach.Preliminary results show promising improvements in the location uncertainty of oceanic seismicity when the proposed data integration approach is applied. Additionally, a refined location catalog, incorporating a more detailed velocity model, is compiled using standard monitoring software alongside the proposed data selection approach.
- 15 May 2025
- Published by Copernicus GmbH
Open AccessA Pragmatic Approach for Integrating Submarine DAS Data with Onshore Seismometer Networks: A Case Study in the Canary Islands
Distributed Acoustic Sensing (DAS) technology offers a valuable opportunity to enhance seismological monitoring of ocean floors. Typically, in a standard monitoring network, most seismic stations are installed on land, with only a few ocean-bottom seismometers available. Consequently, monitoring oceanic seismicity is intrinsically challenging due to the lack of observations near seismic sources. In this context, DAS systems, for instance, those installed on fibers connecting islands and land-islands, can help bridge this observational gap. However, integrating the spatially dense DAS data (meter scale) with the often sparser seismometer data (kilometer scale) is not straightforward. Specifically, inverting DAS alongside seismometer P- and S-wave arrival times can lead to biased location results due to the numerical disparity between datasets and/or outliers not identified. Moreover, employing the full set of DAS arrival times can be computationally intensive, limiting its feasibility for real-time monitoring and integration into routine seismological software.Automated weighting methods can help mitigate bias introduced by arrival time outliers in data inversion. This is particularly useful for DAS data, where user control over individual channels is limited. However, suppose the goal is to use DAS as a complementary tool to a seismometer network, and meter-scale spatial density is not essential. In that case, DAS data selection/sub-sampling can improve computational efficiency. To this end, we propose an approach that, for a given seismological network and DAS system, a) automatically identifies “reliable” DAS channels using a machine learning classifier trained on specific data attributes and b) further selects a subset of DAS channels to achieve similar interchannel spacing to the network. The proposed strategy generates a final set of DAS P- and S-wave arrival times with a number of observations comparable to the network. To test the benefits of this procedure on oceanic seismic monitoring, we use data from a fiber optic cable northeast of Gran Canaria and seismometers operated by the Instituto Geográfico Nacional (IGN) in the Canary Islands. We then compare event locations obtained using: a) IGN-only P- and S-wave arrival times, b) DAS-only P- and S-wave arrival times (unselected), and c) IGN and selected DAS P- and S-wave arrival times using the proposed method. Event locations are estimated using a hierarchical Markov Chain Monte Carlo approach.Preliminary results show promising improvements in the location uncertainty of oceanic seismicity when the proposed data integration approach is applied. Additionally, a refined location catalog, incorporating a more detailed velocity model, is compiled using standard monitoring software alongside the proposed data selection approach.
- Open Access
- 8 April 2025
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 241 (3) , 1923-1935
KVP: a multiscale kurtosis approach for seismic phase picking
SUMMARY: Automatic event detection and phase picking are critical for processing the large volumes of data produced by modern seismological instrumentation. Accurate picking is especially challenging in Distributed Acoustic Sensing (DAS) recordings, where data quality can significantly vary along segments of the fibre due to localized environmental noise and coupling issues, reducing signal-to-noise ratios (SNR). Similarly, Ocean Bottom Seismometer (OBS) data quality also suffers from these issues. To improve accuracy under diverse conditions, we developed a novel multiband kurtosis-based picking algorithm, Kurtosis-Value-Picker (KVP), that enhances phase picking for both impulsive and emergent seismic signals. Our approach uses characteristic functions (CFs) calculated with sliding windows across multiple frequency bands. Triggers are identified based on localized kurtosis jumps over a few samples, providing greater sensitivity to emergent signals than traditional finite-difference methods. Each individual CF has its own set of triggers, adding flexibility to phase picking and retaining spectral information. We validate the KVP algorithm using earthquake data recorded with DAS on two land and submarine fibre-optic cables, as well as OBS data. We also compare its performance with a widely cited, kurtosis-based algorithm, the widely used FilterPicker algorithm and the well-known PhaseNet model, using impulsive signals on nearby DAS channels as a ground truth for emergent arrivals. Our results demonstrate that KVP provides accurate picks and is suitable for complex seismic data sets.
- 8 April 2025
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 241 (3) , 1923-1935
Open AccessKVP: a multiscale kurtosis approach for seismic phase picking
SUMMARY: Automatic event detection and phase picking are critical for processing the large volumes of data produced by modern seismological instrumentation. Accurate picking is especially challenging in Distributed Acoustic Sensing (DAS) recordings, where data quality can significantly vary along segments of the fibre due to localized environmental noise and coupling issues, reducing signal-to-noise ratios (SNR). Similarly, Ocean Bottom Seismometer (OBS) data quality also suffers from these issues. To improve accuracy under diverse conditions, we developed a novel multiband kurtosis-based picking algorithm, Kurtosis-Value-Picker (KVP), that enhances phase picking for both impulsive and emergent seismic signals. Our approach uses characteristic functions (CFs) calculated with sliding windows across multiple frequency bands. Triggers are identified based on localized kurtosis jumps over a few samples, providing greater sensitivity to emergent signals than traditional finite-difference methods. Each individual CF has its own set of triggers, adding flexibility to phase picking and retaining spectral information. We validate the KVP algorithm using earthquake data recorded with DAS on two land and submarine fibre-optic cables, as well as OBS data. We also compare its performance with a widely cited, kurtosis-based algorithm, the widely used FilterPicker algorithm and the well-known PhaseNet model, using impulsive signals on nearby DAS channels as a ground truth for emergent arrivals. Our results demonstrate that KVP provides accurate picks and is suitable for complex seismic data sets.