As a community of geomorphologists, geophysicists, glaciologists, hydrologists, volcanologists, engineers, and ecologists, we advance theory, develop methods, and apply seismic observations to pressing questions in Earth surface research and natural hazards.
We invite contributions on methodological and theoretical developments, field and laboratory experiments, and innovative applications. We particularly welcome work that combines seismic observations with complementary data streams (e.g., remote sensing, in-situ monitoring, fiber-optic networks, or meteorological records), as well as studies leveraging data-intensive approaches (e.g., large-scale arrays, distributed acoustic sensing, machine learning, physics-informed modeling).
We anticipate a lively discussion on current challenges in understanding Earth surface processes, opportunities for community-based research and open data initiatives, and the role of seismic methods in addressing urgent questions related to climate change, natural hazard resilience, and coupled Earth system dynamics.
Topical keywords: erosion, landslide, rockfall, debris flow, granular flow, fracturing, stress, snow avalanche, icequake, calving, subglacial processes, karst, bedload, flood, GLOF, early warning, coastal processes, tsunami, eruption, tremor, turbidity current, groundwater, soil moisture, dv/v, noise, HVSR, array, DAS, infrasound, machine learning, classification, signal processing, physics-informed modeling, multi-sensor integration, open data.
Farming practices reshape soil hydrodynamics by altering near-surface structure, mechanical stiffness, and water transport pathways, yet their impacts remain difficult to observe at field scale and high temporal resolution. Here we combine distributed acoustic sensing with physics-based hydromechanical modeling to quantify how tillage systems and soil compaction influences minute-scale, meter-scale seismic and hydrological responses in agricultural soils. We show that dynamic capillary effects govern transient soil stiffness and moisture redistribution following rainfall, with disturbed soils exhibiting sharp post-rain seismic velocity reductions associated with near-surface saturation. These responses are followed by pronounced hysteretic velocity recoveries driven by evapotranspiration, revealing strong memory effects in soil–water dynamics. Seismically inverted estimates of soil saturation demonstrate how farming-induced disturbance reshapes water flux partitioning and subsurface storage. Our results provide direct observational evidence that farming practices fundamentally reorganize soil hydrodynamics and establish distributed seismic sensing as a scalable, non-invasive approach for observing soil processes relevant to land–atmosphere exchange, Earth system modeling, and resilience to hydrological extremes.
How to cite: Shi, Q., Denolle, M., Montgomery, D., Swann, A., Cristea, N., Williams, E., You, N., Collins, J., Prada Barrio, A., Jeffery, S., Misiewicz, P., and Nissen-Meyer, T.: How farming practices reshape soil hydrodynamics, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16077, https://doi.org/10.5194/egusphere-egu26-16077, 2026.
Soil is a complex ecosystem at the heart of survival for all life on land. Harbouring more carbon than the atmosphere and vegetation combined, it is home to more than 60% of Earth's species and delivers 99% of calories for the human food system. Despite growing demands, more than 70% of global arable land is classified as degraded. Monitoring soils and thereby improving soil health at scale is difficult due to their multiscale heterogeneity, limited accessibility of remote sensing techniques, and destructive, labour-intensive nature of soil coring, on the other hand. Geophysical techniques offer a tangible alternative. To date, active seismics have scarcely been considered for the living topsoil, a layer mere 10-50 cm below our feet.
We show how seismology with ultrahigh frequency wavefields above 500 Hz generated by hammer strikes and recorded by cheap, bespoke geophones should allow us to infer on a variety of crucial soil health parameters, such as bulk density, soil moisture, topsoil depth, soil carbon, and more, which collectively give rise to determining soil function and health. We present consistently high data quality up to 1500 Hz collected across three continents in more than 10 ecosystems and crop types, showcase a pathway for automated data processing and inference, introduce novel low-cost MEMS sensors, and highlight emerging AI engines.
Our non-profit organisation, Earth Rover Program, is tasked with implementing the vision towards a global soil health assessment of unprecedented resolution and coverage. This, in turn, can eventually equip farmers with spatially explicit, local knowledge of their soils’ state and suggest remedial measures based on this novel data, with the potential to reduce environmental pressures and agricultural costs while increasing long-term yields.
How to cite: Bagagli, M., Davidson, K., Tsekhmistrenko, M., Collins, J., Wangechi, M., Mosongo, P., Leng, K., Meng, J., Masson, Y., Jeffery, S., and Nissen-Meyer, T.: A seismic shift for soil health monitoring: scalable, non-invasive seismology at the decimetre scale., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17225, https://doi.org/10.5194/egusphere-egu26-17225, 2026.
Quantifying the source levels and localizing blue whale vocalizations in the Gulf of St. Lawrence is essential for effective management of this endangered population in a region of intense shipping activity. We present an innovative study leveraging data from diverse observational platforms, i.e., ocean bottom seismometers and underwater gliders. Since 2019, gliders equipped with hydrophones have been deployed every summer in the Honguedo Strait shipping lane between Gaspé and Anticosti Island, in the Gulf of St. Lawrence, to trigger mitigation measures when detecting North Atlantic right whales. These platforms have recorded acoustic signals produced by many other whale species, including the endangered Northwest Atlantic (NWA) blue whale, for which similar mitigation measures have not been established. This is because the source level of NWA blue whale vocalizations (e.g. Arch calls) and their detection range from gliders in the region remain unquantified. Yet, these parameters are necessary for cetacean density estimation and for evaluating the feasibility of using this call type for dynamic management frameworks. We take advantage of four ocean bottom seismometers (OBS) on which NWA blue whale calls were detected. We analyse a 1-hour subset of 79 Arch calls that were co-detected on glider and OBS data. The vocalizing whale is localized using a time-difference-of-arrival approach with a coupled ocean-acoustic model that incorporated spatially varying bathymetry, sound speed, and sediment properties. Received levels at each OBS were calibrated using a transfer function, derived from simultaneous particle-velocity and pressure measurements, to quantify the response of the seismometer to the waterborne acoustic wave. Source levels were then estimated using the calibrated received levels and a parabolic equation transmission loss model configured to the environment along the source-receiver path. Preliminary results from our case study demonstrate that Arch calls can be detected in the Northwest Gulf up to 125 km and 150 km from an OBS and glider, respectively, where the difference is primarily explained by the receiver’s position in the water column. The measured acoustic features of Arch calls suggest propagation similarities to infrasonic blue whale vocalizations (i.e., songs). These findings have ecological implications and can inform management strategies in an area heavily used by both whales and vessels.
How to cite: Goblot, E., Gehrmann, R., Barclay, D., Indeck, K., Plourde, A., Sirati, E., and Nedimović, M.: Seismoacoustic Localization and Source Level Estimation of Blue Whale Calls in the Gulf of St. Lawrence, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11529, https://doi.org/10.5194/egusphere-egu26-11529, 2026.
On 10 August 2025, a large landslide (>64×10⁶ m³) collapsed more than 1,000 m onto South Sawyer Glacier and into Tracy Arm fjord in Southeast Alaska. The resulting tsunami ran up the opposing fjord wall to a height of 480 m, the second-highest tsunami ever recorded.
The landslide was preceded by more than 24 hours of repeated microseismicity (up to M~2), with event rates increasing until ~1 hour before failure, signalling a transition to continuous slip of the overall rock mass.
The landslide generated globally observed body and long-period seismic waves equivalent to an Mw 5.4 earthquake, making it one of the largest-magnitude landslides in decades. From regional and global seismic data, we infer a total mass of ~370 million metric tons, exceeding estimates from remote sensing and DEM analysis. This discrepancy suggests that water displacement during the initial tsunami contributed to the long-period global seismic signal.
Following the landslide signal, we observe monochromatic seismic waves worldwide with dominant periods of 50, 52, 66, and 86 s. The 66 s mode is strongest and persists for >36 hours at regional stations. Surface-wave radiation patterns, numerical tsunami modelling, and SWOT satellite water-height observations support the genesis of a fjord-transverse landslide-induced seiche (LIS) in the central fjord. However, seismic radiation is more complex than that of the 2023 Dickson Fjord, Greenland LIS event, likely reflecting differences in the landslide location and its direction, fjord geometry, and interaction of multiple seiche modes.
Despite heavy summer vessel traffic in Tracy Arm, there were no fatalities, making this a near miss. Seismic observations, combined with remote sensing, provide a critical pathway for forecasting and early warning of cascading landslide–tsunami events and for understanding ice–land–water interactions in polar environments.
How to cite: Hicks, S., Shugar, D., Berdahl, M., Caplan-Auerbach, J., Ekström, G., Fathian, A., Geertsema, M., Higman, B., Karasozen, E., Lynett, P., Monahan, T., Roe, G., Svennevig, K., Van Wyk de Vries, M., and West, M.: Seismic observations and modelling of the August 2025 Tracy Arm, Alaska landslide, megatsunami, precursory seismicity, and seiche, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17665, https://doi.org/10.5194/egusphere-egu26-17665, 2026.
Large earthquakes can trigger cascading environmental impacts in continent (e.g. 1964 Mw 9.2 Alaska and 2015 Mw 7.9 Nepal earthquakes), yet such processes remain poorly documented in oceanic and polar settings. Here, we present a multidisciplinary investigation of the 2025 Mw 6.5 oceanic strike-slip earthquake that occurred on 10 March 2025 along the Jan Mayen oceanic transform fault, Arctic Ocean, and its surface and cryospheric impacts on local Jan Mayen Island. Using relocated local seismicity, regional waveform modelling, GNSS time series, seismic noise interferometry, infrasound observations, high-resolution optical satellite imagery, and long-term air-temperature records, we reconstruct the sequence of events linking the earthquake rupture to a major rock-slope failure. The earthquake, which ruptured for ~40 km long in transform faulting, triggered a rock avalanche from a steep, glacier-adjacent volcanic slope, depositing 0.8-1.2x106 m3 of debris material over ~0.9 km2 of the Kjerulf Glacier and reaching the coastline. Infrasound signals constrain the timing of slope failure to within minutes of the mainshock, supporting a co-seismic trigger. Satellite imagery further reveals contemporaneous calving at the Weyprecht Glacier. We observed from pre-2025 satellite imagery that multiple smaller rock-slope failures between 2019 and 2024, indicating progressive slope weakening prior to the earthquake. Also, long-term air-temperature records show a marked warming trend over recent decades, including the near absence of extreme cold winters since the late 1990s and an increasing frequency of anomalously warm summer days, consistent with Arctic amplification. We interpret the 2025 rock failure at Kjerulf Glacier as an earthquake-triggered collapse of a slope preconditioned by permafrost degradation associated with this warming trend. Our results demonstrate that oceanic strike-slip earthquakes can generate significant onshore geohazards in polar environments and highlight the importance of integrated geophysical and remote-sensing approaches for monitoring earthquake-climate-cryosphere interactions in the Arctic.
How to cite: de Melo, G. W. S., Hermanns, R., Bendle, J. M., Grevemeyer, I., Cesca, S., do Nascimento, A. F., Ottemöller, L., Aslan, G., Brissaud, Q., Oye, V., and Kopp, H.: An earthquake-triggered rock avalanche on Jan Mayen Island conditioned by Arctic warming, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-22884, https://doi.org/10.5194/egusphere-egu26-22884, 2026.
Using a new global seismological analysis technique designed to detect long-lasting coherent signals, we identify both previously known and entirely unrecognized resonant-like seismic emissions at periods longer than 10 seconds. A detailed examination of these signals allows us to locate their sources with remarkable precision. Strikingly, they cluster in offshore sedimentary basins near major river fans and beneath ice-covered regions. Although their resonant character resembles classic volcanic tremor, the source locations indicate that they are not associated with any known volcanic system.
A careful analysis of their frequency content, spatial distribution, and radiation patterns instead suggests that these signals may originate from the resonance of fluids within shallow subsurface reservoirs. This interpretation aligns with the presence of large volumes of gas, oil, and water in thick sedimentary basins, and with seafloor seepage structures that release substantial amounts of naturally generated fluids from depths of roughly 5–10 km.
By tracking the temporal evolution of these signals, we also identify a pronounced seasonal modulation that mirrors oceanic variability. This observation points to a significant coupling between the oceans and the solid Earth, potentially mediated by static or dynamic stress transfer.
The detection of these newly recognized signals opens a promising path toward probing the largely unexplored dynamics of sedimentary layers and their sensitivity to external environmental forcing. More broadly, these findings introduce a new class of geophysical observables capable of revealing how the shallow lithosphere responds to, and interacts with, oceanic processes on seasonal to long-term timescales.
How to cite: Poli, P. and Takano, T.: Detection and characterization of resonant signals in global seismology: Evidence for shallow fluid reservoirs and their interaction with the oceans, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1453, https://doi.org/10.5194/egusphere-egu26-1453, 2026.
Monitoring dynamic surface processes such as sediment transport and turbulence during high-flow events remains a major challenge in fluvial geomorphology. Seismic methods provide a non-invasive alternative, but robust source localisation is challenged by the distributed nature of fluvial sources, heterogeneous shallow structures, and the broadband character of the signals they produce. Because turbulence and bedload dominate different parts of the spectrum, we require broadband analysis. We establish a generalised framework for frequency-resolved seismic source localisation using dense arrays and matched field processing. We introduce a hybrid processing strategy that exploits the array differently across frequencies: full-network matched field processing at low frequencies, where coherence spans the entire aperture, and sub-array averaging at higher frequencies, where coherence is confined to local scales. We apply the framework to a field case, where we retrieve frequency-dependent source regions across the active channel and separate low-frequency turbulent noise from higher-frequency bedload impacts. We conduct synthetic tests to quantify localisation uncertainty as a function of frequency, sensor density, and signal-to-noise ratio. Across the 2-40 Hz range, we find that localisation uncertainty varies from a few metres at low frequencies to more than 100 m at high frequencies, reflecting the expected loss of resolution at shorter wavelengths. By quantifying these trade-offs, we provide practical guidance for future deployments, including sensor spacing and array geometry required to achieve a target resolution.
How to cite: Venkatesh Revankar, S., Gimbert, F., and Recking, A.: Frequency-Resolved Seismic Source Localization in Fluvial Settings Using Dense Arrays, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4482, https://doi.org/10.5194/egusphere-egu26-4482, 2026.
The manner by which glaciers evacuate the products of erosion is poorly known, especially for the coarser-sized sediments produced by plucking and quarrying, that is the bedload fraction. It has traditionally been assumed that bedload export from Alpine glaciers is relatively efficient. However, continuous, seismic measurements of bedload transport at glacier portals question this assumption as do the results of numerical modelling experiments. There are a number of possible hypotheses to explain such inefficiency, including the structure and geometry of subglacial channels, periodic variation in discharge due to diurnal melt cycles and near-ice margin effects. The problem remains, however, that there are no published datasets on subglacial bedload transport to investigate the effects, not surprising because of the challenges associated with measuring it. The aim of this research is to harness environmental seismology to quantify subglacial bedload transport rates for the first time using on-ice measurements.
In the first part of the project, we traced the location and development of two subglacial channels of Glacier d’Otemma, Valais, Switzerland. In this study, we investigate the discharge and sediment traveling in our identified channels. With the help of thirty seismic sensors on the glacier surface and at the glacial terminus, we record the seismic activity of Glacier d’Otemma during its melt season in 2024. We trace the path the sediment takes underneath the glacier with seismic beamforming techniques and relate it to the previously identified shape of the subglacial drainage system. A line of seismic sensors along the glacier was able to record waves of bedload as they travel underneath the ice. With the help of physical models, the spectral information of the ground motion data is translated into sediment transport and discharge quantities. The project contributes to reveal the interaction of sediment and the characteristics of subglacial drainage system in alpine glaciers, giving further insights into erosion mechanisms at play.
How to cite: Wolf, E., Mancini, D., Dietze, M., Stutzmann, E., Metaxian, J.-P., and Lane, S.: Observing subglacial bedload transport dynamics with on-ice seismic networks on Glacier d’Otemma, Switzerland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9649, https://doi.org/10.5194/egusphere-egu26-9649, 2026.
Long-term seismic monitoring provides a unique insight into glacier and ice-shelf dynamics. However, the extraction of meaningful cryoseismic information from continuous multi-year records remains challenging. Icequake events frequently show unclear or overlapping signals due to harsh environmental conditions and persistent background activity. While the utilisation of variable event lengths can be instrumental in the avoidance of merging multiple events into a single window, most unsupervised learning methods require fixed input durations. This emphasises the necessity for a novel unsupervised clustering approach that can handle time-variant events of varying lengths while robustly detecting outliers. The method should be physics-based to accommodate the limited prior knowledge of icequake characteristics. It should also operate directly on large event catalogues, without reliance on handcrafted features.
To address this issue, the incrementally buffered dynamic time warping clustering is introduced. This is a new approach to clustering dynamic time warping (DTW) distances of events and it incorporates the requirements stated above. The method starts with an initial k-medoids clustering on a pairwise DTW distance matrix of a subset of events, thereby generating initial k clusters. The subsequent addition of new samples is based on a statistically robust distance threshold from the distribution in within-cluster distances of the initial step. Each event is compared only to existing medoids, assigned to the nearest cluster if the DTW distance falls below the threshold, or temporarily placed in a buffer when classified as an outlier. The promotion of buffered events to new clusters is only permitted when the criteria for similarity and minimum sample count are met, thus preventing the formation of spurious clusters from isolated noise events. Lastly, a final global reassignment step is performed. This step involves the recomputation of all event-to-medoid distances. The purpose of this is to stabilise cluster boundaries and refine the catalogue. The combination of these steps results in a scalable and transparent algorithm that is well-suited to the analysis of extensive environmental time-series data.
The present study applies this framework to a time span of several years of vertical-component seismic data from the 16-sensor Watzmann array at Neumayer Station III, Antarctica. Preliminary results indicate the presence of numerous persistent families of icequakes. These are analysed with regard to their correlation with environmental conditions, including tidal modulation and wind.
How to cite: Kiel, A., Hammer, C., and Schlindwein, V.: A Robust Framework for Clustering Variable-Length Seismic Events: A Cryogenic Case Study , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9781, https://doi.org/10.5194/egusphere-egu26-9781, 2026.
Rock fracturing plays a key role in the formation of mountain landscapes and natural hazards. Freezing is one of the main triggers of erosion and fracturing in the Alps. However, questions remain about the impact of freezing on the resonance frequency of rock pillars and the quantification of mechanical stress generated by ice in natural cliffs.
To better understand the effect of frost at the centimeter and meter scale, long-term recordings were made using repeatable ultrasonic signals to measure both sound velocity and waveform changes. These observations are combined with the measurement of the pillar's fundamental resonance frequency. The investigated Tête Noire rock pillar consists of micaschist and is instably hanging above the city of Trient in the western Swiss Alps.
The results show that during freezing, the fundamental resonance frequency increases by 50 %, the P-wave velocity increases by 17 %, and for later arrivals (coda wave) velocity increases by 4 %. After the freezing period, a irreversible drop in P-wave and coda wave velocities is visible, but not in the fundamental resonance frequency which is coming back to its initial value. This decrease in velocity is accompanied by a decorrelation of the ultrasonic waveforms. Reproducing the observed P wave velocity changes on 0.5m thick layer on a finite element COMSOL simulations of the pillar, we determine that the changes in velocity in the rock do not explain the fundamental resonance frequency changes. We therefore propose that the increase in fundamental resonance frequency results from ice filling the rear crack, and we estimate an order of magnitude of about 1.7 m for the ice height, compared with the initial crack size of 10 m.
To estimate the freezing stress, from the measured velocity changes, we determined the acousto-elastic constant of the Tête Noire micaschist on a representative laboratory sample using uniaxial compression experiments. Those results reveal that the generated freezing induced stress are in the subcritical regime with an order of magnitude of a few tens MPa and are, hence able to damage slightly the rock, irreversibly. The drop in correlation coefficient and in the waves velocity support this conclusion.
This work was funded by the European Research Council (ERC) under grant No. 101142154 - Crack The Rock project
How to cite: Rousseau, R., Starke, J., Bottelin, P., Moreau, L., Baillet, L., and Larose, E.: Estimation of Frost-Induced Stress Using Time-Lapse Ultrasonic Testing and their Effect on the Dynamics of an Alpine Rock Pillar, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17386, https://doi.org/10.5194/egusphere-egu26-17386, 2026.
In mountain catchments, floods can mobilize large amounts of sediment, yet monitoring these events remains a major challenge. Understanding the processes governing sediment transport during extreme floods, such as flood waves and sediment pulses, is key to improving our understanding of rapid erosion dynamics. Environmental seismology offers a powerful approach to detect and quantify these processes remotely and continuously with high temporal resolution.
The Mediterranean basin is characterized by a climate and topography prone to flash floods. The objective of this work is to quantify the sediment transport that occurred during the extreme flooding associated with Storm Alex (October 2020), and the subsequent major flood caused by Storm Aline (October 2023), on the Roya River in southeast France.
To address this objective, we use seismic measurements from a single-component geophone (natural frequency of 4.5 Hz) installed at 5 m from the Roya River. We apply previously developed physical models describing the seismic power generated by river bedload transport (saltation model) and turbulent flow. Comparison with model predictions suggests that, at such short distances, the recorded seismic power is dominated by the bedload process, allowing us to focus on the saltation model.
To quantify the sediment transport during periods of peak seismic amplitude, we calibrate the parameters of the saltation model to the Roya River context. Although the resulting volumetric sediment flux is consistent in order of magnitude with theoretical and empirical estimates, the complexity of the physical environment calls for further investigation. Seismic parameters of the riverbed, realistic grain size distributions of transported sediments, and local hydrometric data remain difficult to constrain directly. We address these uncertainties through sensitivity analyses, which show that seismic medium parameters mainly control the shape of the seismic spectrum. We therefore explore the use of real data that we obtain from an active and passive seismic experiment to adjust those parameters
Water depth is another key parameter of the saltation model, as it controls the basal shear stress. We estimate flow depth using upstream discharge measurements and local discharge modeling with simplified theoretical relationships between flow depth, river width, and discharge. Ongoing and future work includes seismic array measurements and local hydraulic modelling to further constrain model parameters. Overall, our results highlight the potential of environmental seismology to quantify sediment transport during extreme flash floods and to improve process-based understanding of sediment transfer in steep Mediterranean river systems.
How to cite: Dib, M.-O., Chmiel, M., Chapuis, M., Ampuero, J.-P., Abily, M., Mercerat, D., and Courboulex, F.: Seismic monitoring of sediment transport during flash floods: Case studies of Storms Alex and Aline on the Roya river, France, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9430, https://doi.org/10.5194/egusphere-egu26-9430, 2026.
High-frequency acoustic and seismic signals generated by granular flows, such as bedload transport and debris flows, provide valuable information on sediment dynamics, yet the physical interpretation remains challenging. In dense or partially dense solid-fluid granular flows, signal generation is controlled not only by particle-bed impacts but also by frequent inter-particle collisions within the actively shearing layer. These collisions are shear-driven and evolve rapidly in time and space, with impact rates being highly sensitive to shear strain rate, time, and granular layer thickness. However, most existing particle impact rate models assume stationary conditions and neglect the spatiotemporal variability inherent in natural geophysical flows, limiting the ability to explain observed non-stationary spectral signatures. Here, we develop a new analytical framework for particle impact rate in solid-fluid two-phase granular flows based on non-equilibrium thermodynamics. The model explicitly links collision rate to shear strain rate, granular state variables, and the thickness of the basal shearing layer, allowing impact rates to evolve dynamically in time and space. Reformulating the model in the frequency domain provides a direct theoretical connection between evolving collision rates and the spectral properties of the generated acoustic and seismic signals. For saturated channel beds, we further investigate the two-way coupling between pore water pressure and particle impacts in signal generation. Particle impacts are conceptualized as transient mechanical sources that locally compact the granular skeleton, reduce pore volume, and generate excess pore pressure, which in turn feeds back on particle impacts. Analytical solutions demonstrate that the amplitude and persistence of impact-induced pore pressure perturbations are controlled by bed permeability, shear strain rate, and the thickness of the basal shear layer. An increase in pore pressure reduces effective stress and feeds back on collision dynamics, introducing an additional control on signal generation. Building on these results, we extend existing power spectral density formulations to show that temporally evolving particle impact rates modulate frequency spectra by redistributing spectral power across frequences, resulting in departures from classical spectral scaling. Pore pressure effects further modify spectral amplitudes and attenuation. The proposed framework offers new physical insights into sediment-generated signals, enabling improved interpretation of evolving particle impact rates and pore pressure related effects in bedload transport and debris flows.
How to cite: Chen, Z., Rickenmann, D., Walter, F., McArdell, B., Kang, J., Wetter, C., and Badoux, A.: Spatiotemporally evolving particle impact rates in sediment-generated acoustic signals, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5278, https://doi.org/10.5194/egusphere-egu26-5278, 2026.
Atmospheric effects, specifically wind and precipitation, are often considered as dominant sources of noise in seismic records, both in the high frequency and ultra low frequency domain. However, those two processes are also important drivers of geomorphic activity: wind causing advective processes, aeolian transport and erosion, tree uprooting, and energy transfer from the atmosphere to the ground. Precipitation, specifically rain, controls splash erosion, surface sealing through puddling, soil moisture, groundwater fluctuations, and importantly, the initiation of surface flow and resulting flood and sediment transport waves. Despite their widespread reflection in seismic data sets, especially in the context of environmental seismology studies, there is surprisingly little valorisation of the signal content associated with these two environmental variables. Here, I show case study based examples of the different signatures of wind and rain in seismic data sets to illustrate the systematic variability arising from different process modes, intensities and types of interaction with elements of the Critical Zone. I make use of physical models of rainfall and modified models of wind interaction with open ground and trees, to explore the information that can be extracted from seismic data sets by inverting those models in combination.
How to cite: Dietze, M.: Surveying and modelling the distributed effects of wind and rain, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10606, https://doi.org/10.5194/egusphere-egu26-10606, 2026.
Surge-type debris flows advance through successive surges, during which in-channel deposition layers progressively develop between surges and continuously modify basal conditions. Seismic observations from Jiangjia Ravine show that ground vibration amplitudes systematically weaken as surge sequences evolve, even when successive surges exhibit similar flow magnitudes, implying a breakdown in the conventional scaling between flow intensity and seismic response. This phenomenon is interpreted as a consequence of the progressive buildup and partial liquefaction of inter-surge deposition layers, rather than the influence of static, pre-existing bed deposits. To represent this process quantitatively, we introduce an effective transmission parameter (ξ) into a fluvial seismology framework and establish a sigmoid relationship between ξ and the normalized thickness of the deposition layer (H*). Incorporating this relationship substantially enhances the ability to reproduce observed variations in seismic power spectral density (PSD) across surge sequences and offers a transferable means of capturing subsurface flow–bed coupling. These results highlight the importance of dynamic bed evolution in controlling debris-flow-generated seismic signals and provide new insights for improving real-time monitoring and early-warning strategies in sediment-laden mountain catchments.
How to cite: Jiang, F. and Song, D.: Liquefied Deposition Layers Modulate Seismic Wave Propagation in Surge-type Debris Flows, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9446, https://doi.org/10.5194/egusphere-egu26-9446, 2026.
The unstable rock slope "Spitze Stei" (Kandersteg, Switzerland) has shown significantly increased activity for several years. Since 2018, observed displacement rates can exceed 40 cm per day seasonally. The instability covers a total area of approximately half a square km. The volume of the moving rock and debris mass is ~16 million m3, distributed across several rock compartments. Driven by degrading permafrost and enhanced gliding planes, these primary gravitational instabilities result in secondary, often destructive, debris flows into the Oeschibach channel. While continuous monitoring is essential for risk management, traditional visual and radar methods are often constrained by adverse weather conditions, limited temporal resolution and limited sensitivity to subsurface processes. To overcome these limitations and monitor rockslide internal deformation, material damage, and ongoing mass-movement processes at high spatial resolution, a dense temporary seismic network consisting of 64 SmartSolo nodes (natural frequency 5 Hz) were deployed across the slope at the end of June 2025 and operated for nearly three months. This dataset is complemented by recordings from three semi-permanent seismometers that have been operating since October 2021, providing a longer-term reference for background seismicity and site-specific noise characteristics.
We analyze the continuous seismic records to detect and characterize signals from a variety of mass-movement phenomena, including rockfalls, granular flows, debris flows and avalanche-related activity. Signals are evaluated based on waveform properties, duration, amplitude evolution, and spectral content, with comparisons across sensor types and periods. A key objective is to isolate and cluster internal microseismic activity, distinguishing it from background noise, external sources (e.g., icequakes), and transient permafrost-related signals.
Our preliminary results highlight a diverse set of seismic signal types linked to both surface processes and internal rockslide dynamics. This observed variability suggests changes in deformation style across different rock compartments, demonstrating the potential of dense nodal seismic arrays to resolve internal rockslide processes relevant for hazard monitoring.
How to cite: Das, B., Chmiel, M., Courboulex, F., Walter, F., Martin, X., Osorno-Bolívar, J.-S., Kienholz, C., Arias, G., and van den Ende, M.: Seismic Transients of Internal Deformation in an Active Rockslide (Spitze Stei, Switzerland): First Insights from a Dense Seismic Nodal Array , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12096, https://doi.org/10.5194/egusphere-egu26-12096, 2026.
Iceberg calving and near-shore landslides in Greenland produces seiches in the fjords - standing tsunami-like waves on the order of minutes of period that resonate for hours (or even days in extreme cases, see Svennevig et al., 2024), and can pose danger to the population and cause damage to infrastructure in the villages. For more than a decade, it has been known that broadband seismic sensors on-shore are sensitive to the ground tilt induced by these waves (Amundsen et al., 2012). Seiches are typically seen in seismic data as very long period waves that last tens of minutes to hours. This means that a network of on-shore seismic sensors can be employed to provide a tsunami early warning (TEW) system for both on-site and network-wide TEW. A major advantage of seismic networks over pressure gauges is the fact that the sensors are not exposed to the destructive forces of sea ice and icebergs, which are abundant in many regions of Greenland. In this work we demonstrate how a seismic network in the Uummannaq fjord (Greenland) can be used to provide TEW to the villages in the fjord. We further demonstrate an algorithm that allows detecting seiches and discriminating them from other sources of long-period signals (mostly large teleseismic earthquakes). Such an algorithm, however, can be significantly affected by corrupted data (spikes, steps, etc.), which produce false alarms. We then demonstrate how we can remove these false triggers using a deep scattering network (Seydoux et al., 2020). Our results show that we can detect seiches with little false alarms and provide timely TEW in the Uummannaq fjord, including the 2017 Nuugatsiaq landslide. We also demonstrate our implementation of the developed TEW algorithm into a real-time system in SeisComP.
How to cite: Jozinović, D., Clinton, J., Massin, F., Seydoux, L., Mätzler, E., and Petersen, J.: Building a Demonstration Tsunami alert system in the Uummannaq fjord, Greenland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12146, https://doi.org/10.5194/egusphere-egu26-12146, 2026.
Proglacial streams that drain Alpine glaciers are characterized by variable hydrological forcing and highly intermittent sediment transport, making continuous monitoring of hydro-sedimentary processes necessary to understand their dynamics. Near-field seismic monitoring has recently been established as a valid non-invasive approach to studying river dynamics, since ground vibrations are sensitive to both flow turbulence and bedload transport.
In this study, we analyse the seismic response of a proglacial stream fed by the Rutor Glacier, the sixth-largest glacier in the Italian Alps, integrating passive seismic monitoring with hydrological and climatic observations. Three geophones were installed close to the active channel and continuously recorded ground vibrations at 200 Hz during the 2025 ablation season (June–September). The seismic power spectral density was analysed across different frequency bands. Water level was monitored using pressure sensors, while discharge was estimated using saline dilution tests, allowing the relationship between seismic signals and hydrological forcing to be investigated.
The preliminary results show marked diurnal fluctuations in water level driven by glacial melt. At low frequencies (5–15 Hz), seismic power increases predominantly linearly with water level, suggesting a dominant control by water flow turbulence. In contrast, at relatively high frequencies (30–40 Hz), the seismic response becomes nonlinear and exhibits a clear change in slope when a critical water level is exceeded, suggesting the activation and/or intensification of bedload transport superimposed on the hydraulic signal.
This study highlights the potential of environmental seismology as a non-invasive and continuous monitoring approach to investigate hydro-sedimentary dynamics in highly variable proglacial environments.
How to cite: Marini, F., Piantini, M., Comiti, F., Bertagni, M., and Camporeale, C.: Frequency-dependent seismic response to hydrological and bedload forcing in a glacier-fed stream, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12535, https://doi.org/10.5194/egusphere-egu26-12535, 2026.
Continuous seismic monitoring has proven effective for detecting rockfalls, yet most studies rely on multiple stations or dense arrays, increasing cost and complexity. This study demonstrates that a single seismic station, located approximately 100 m from the event site, can detect small rockfalls (<0.005 m³), characterize their dynamics, and estimate their volumes.
The approach relies on careful signal processing, combining STA/LTA analysis, envelope calculation, and parameters such as amplitude, duration, and frequency, to reveal distinctive features of rockfall events. This methodology emphasizes the quality of extracted information over the quantity of data, enabling real-time identification of minor precursory events even amidst diverse environmental and anthropogenic noise.
LiDAR and photogrammetry provide high-resolution spatial data to calibrate and validate detections, but their limited temporal resolution prevents continuous monitoring. Controlled block-fall experiments further optimized station placement and confirmed the system’s sensitivity. These results demonstrate the potential of cost-effective, single-station seismic monitoring for automatic rockfall detection and early warning, offering a practical solution for hazardous mountainous regions.
How to cite: Tapia, M., Guinau, M., Blanch, X., Abellan, A., Telletxea, B., Martín, J., and Meneses, F.: Seismic Monitoring of Rockfalls Enhanced by LiDAR and Photogrammetric Data: Towards Automatic Detection and Early Warning., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16328, https://doi.org/10.5194/egusphere-egu26-16328, 2026.
Information about frozen lakes, including ice rigidity, ice thickness, and water depth, is essential for environmental studies and practical applications. Although these properties can be measured in the field, such measurements are labor-intensive and spatially limited, motivating the development of alternative observation methods. Seismic waves provide an effective approach to studying frozen lakes, as their propagation velocity depends on the physical properties of the ice–water system, including the elastic moduli and thickness of the ice, and water column depth. In this study, we investigate the use of wind-driven flexural waves recorded by a distributed acoustic sensing (DAS) system to infer ice thickness and water depth beneath a 1000 m fiber-optic cable installed on Lake Pääjärvi, southern Finland. We identify wind-induced flexural waves in the 0.01-0.5 Hz frequency band, extract their dispersion curves, and invert them using a grid search to estimate effective ice thickness and water depth under four cable intervals. Our estimates indicate effective ice thicknesses ranging from 22 to 34 cm and effective water depths ranging from 0.8 to 31 m. Absolute differences between effective estimates and arithmetic averages of field measurements range from 0.5 to 8.6 cm for ice thickness and 1.2 to 10.3 m for water depth. Our estimates reproduce the observed dispersion curves and agree with field measurements, demonstrating that it is possible to obtain first-order information about ice thickness and water depth in frozen lake environments. However, the robustness of water depth estimates is limited by the wavenumber content of the flexural waves. In our case, the uncertainty of the water depth estimates increases from 0.43 to 12.05 m as water depth increases because low-wavenumber flexural waves, which are most sensitive to the water column, are not resolved by the dispersion curves. Another important observation is that refraction of flexural waves toward shallower water must be considered when converting apparent velocities measured along the cable to true velocities. If this effect is neglected, dispersion curves and the estimated parameters can be biased.
How to cite: Valero Cano, E., Moreau, L., Strobel, F., and Hillers, G.: Using flexural waves recorded by distributed acoustic sensing to infer the ice thickness and water depth of a frozen lake, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16347, https://doi.org/10.5194/egusphere-egu26-16347, 2026.
Environmental seismology increasingly employs relative changes in seismic velocity (dv/v) with time, derived from continuous seismometer recordings, to infer temporal variations in ground properties such as soil moisture and groundwater dynamics. However, despite that widespread use, there is no consensus on optimal frequency bands and stretching time windows used to extract reliable dv/v time series. Previous studies addressing deep, and shallow groundwater dynamics each applied distinct combinations of frequency bands and stretching time windows, raising the key question: how comparable are dv/v results derived from different parameter choices, and how consistently do they represent subsurface hydrological variability?
This work presents an intercomparison of commonly used noise-based dv/v combinations of frequency bands and stretching time windows across two hydrological domains: (1) groundwater at depth, and (2) shallow critical zone hydrology. For each domain, we review and implement published combinations of the previous parameters and assess their influence on the resulting dv/v time series.
To evaluate their methodological impact, we compare the different dv/v estimates amongst themselves and against independent environmental control datasets, including groundwater levels, soil moisture time series, and additional hydroclimatic observations. For both domains, we complement literature case studies with our own two field datasets from Germany, one in the Eifel and one in the Harz region. In the Eifel region, a network of ten seismic stations has been deployed across two sub-catchments in the upper Ahr Valley, composed of intensively folded and fractured Devonian mudstone and carbonate rocks. Continuous seismic records are paired soil moisture sensors at depths down to 40 cm and groundwater well measurements from nearby monitoring sites. In the Harz Mountains, we use additional seismic, top soil moisture, and hydrological observations to extend the comparison to a geologically and hydroclimatically distinct setting, with variable soil cover on weathered granite boulders creating abundant water flow in the shallow subsurface.
We hypothesize that different combinations of frequency bands and stretching time windows will produce systematically distinct dv/v patterns, even when applied to the same dataset, and that their sensitivity will vary across deep groundwater systems and shallow surface hydrology. By identifying parameter combinations that enhance or mask relations with the independently sensed environmental variables, this study aims to better understand methodological controls between these parameters and contribute towards a more consistent and comparable practice for environmental applications.
How to cite: Fracica Gonzalez, L. R., Andermann, C., Abadie, B., Armitage, J., Dietze, E., Hovius, N., Illien, L., Putzenlechner, B., and Dietze, M.: Evaluating Frequency Band and Stretching Time Window Effects on Noise-Based dv/v for Subsurface Hydrological Monitoring, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18499, https://doi.org/10.5194/egusphere-egu26-18499, 2026.
The composition, structure, and dynamics of a transient ice sheet that forms and disintegrates on a boreal lake is influenced by meteorological and environmental processes. This includes trapping of upwelling methane from the lake sediments, which is in turn affected by eutrophication in the catchment area. Methane is a potent greenhouse gas, yet documented sources and sinks to the atmospheric budget are highly unbalanced. Here we explore a novel approach for quantifying methane ebullition from a boreal lake that combines seismic methods together with interdisciplinary observation methods.
The DYNALake project centerpiece is an array of ~210 seismic geophones arranged in an aperiodic tiling configuration that we deployed in February 2025 on the ~20 cm thick ice of Lake Pääjärvi some 100 km north of Helsinki. The 10-km scale lake array is complemented by a sparser network of 31 land-based sensors installed around the lake between fall 2024 and spring 2025, three dense circular arrays enabling local beamforming and estimating array derived rotation, a DAS system with a 1 km-long fibre optic cable, an underwater echosounder to monitor potential methane ebullition, a rotational seismometer, a microphone to record seismo-acoustic waves, a Ground Penetrating Radar (GPR) survey, water chemistry measurements, manual ice thickness sampling and ice coring, and meteorological data. The project popularizes the subarctic wintertime fieldwork and the science by making a professional documentary for science communication, outreach, and education.
We present initial results on spatial and temporal variations in lake-ice thickness and on the quality and characteristics of the recorded seismic data. The observations include distinct ice-guided wavefield signatures, including QS₀ (quasi-symmetric) and SH₀ (horizontally polarized shear) modes used to estimate elastic parameters such as Young’s modulus and Poisson’s ratio, as well as the dispersive QS (quasi-Scholte) mode that is primarily sensitive to ice thickness at higher frequencies. We compare signals from natural sources and hammer shots across the different sensor types. We show examples of noise correlation wavefields, beamforming results, and seismo-acoustic records that can be used to characterize seismic activity patterns and resolve variable ice properties. Seismic activity in the 0.03–0.2 Hz band increases during high-wind episodes, while higher-frequency signals (0.1–1000 Hz) correlate with rapid air-temperature cooling events. The GPR profile images the spatial ice variability across the lake that is compatible with the in situ measurements. The geochemical water sample analysis suggests Lake Pääjärvi is a source of methane.
We discuss the potential of the data quality and the sensor configuration for signal detection and for icequake and passive tomography lake ice images to resolve spatially variable air and gas bubble properties that are controlled by environmental processes. This synthesis demonstrates that the application of environmental seismology concepts can form a bridge between bottom-up ebullition monitoring and remote-sensing approaches.
How to cite: Strobel, F., Hillers, G., Jilbert, T., Loehr, J., Stranne, C., Oksanen, T., Vänskä, J., Courbis, R., Rintamäki, A., Sadeghi-Bagherabadi, A., Weißgräber, L., Ding, Y., de Langenhagen, M., Valero Cano, E., Mordret, A., Schmelzbach, C., Moreau, L., Coutant, O., and Hadziioannou, C. and the DYNALake deployment team: Resolving environmental processes by imaging and monitoring lake ice properties of the boreal Lake Pääjärvi, southern Finland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18667, https://doi.org/10.5194/egusphere-egu26-18667, 2026.
Landslides remain one of the most pervasive natural hazards in the Himalayan region, exacerbated by intense rainfall, steep topography, and anthropogenic activity. Accurate detection, monitoring, and hazard zoning of these slope failures are essential for mitigating their impact on communities and infrastructure. This research integrates Synthetic Aperture Radar Interferometry (InSAR) with Seismic Ambient Noise Interferometry (SANI) monitoring to provide a comprehensive assessment of landslide-prone regions. Using time-series InSAR techniques such as Persistent Scatterer (PS-InSAR) and Small Baseline Subset (SBAS) interferometry, implemented through open-source tools like OpenSAR Lab, MintPy, LiCSBAS, and StaMPS, we processed multi-temporal SAR datasets from January 2020 to December 2024 to retrieve surface deformation rates. This analysis enabled the identification of both active and slowly deforming landslides across the study area.
The seismic velocity change was computed to monitor subsurface behavior from 17th June 2024 to 17th July 2024 using three seismic stations, viz AULI, CHAD, and KVSJ, located in Joshimath, Chamoli. The dv/v calculate for KVSJ shows a prominent velocity drop on July 5th, 2024. The landslides occurred on the 9th and 10th of July 2024. The drop in velocity can be attributed to increased moisture due to precipitation, which results in rigidity loss. The observed velocity drop may be interpreted as a precursor signal to landslide occurrence. However, the dv/v plots for the AULI and CHAD do not show any prominent drop in the velocity. The one possible reason could be the aspect of the downslope area that orients to a relatively stable valley bottom. As the region hosts metasedimentary rock, the subsurface response to the infiltration of rainwater could also have contributed to the different behavior of these two stations, and this is to be further investigated to identify the plausible causes.
The integrated methodology presented in this work demonstrates that the synergy between InSAR and passive seismology not only improves landslide detection and monitoring capabilities but also contributes to more informed early warning systems and risk reduction strategies. This study contributes to the broader goal of disaster-resilient infrastructure planning in mountainous terrains.
How to cite: Mishra, S. and Maurya, S.: Landslide Monitoring in Joshimath through Passive Seismology and SAR Interferometry, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21031, https://doi.org/10.5194/egusphere-egu26-21031, 2026.
Near-surface groundwater dynamics (NSGD) reflect residence and recharge of the terrestrial water and control the water resource of the downstream area. Environmental seismology, using seismic velocity changes (dv/v), provides a non-invasive approach to observe NSGD. Here, we use ambient seismic noise at four seismic stations, hydraulic, and meteorological records from the Liwu watershed, Taiwan, following Typhoon Kompasu in 2021 to investigate NSGD in response to typhoon rainfall. We used the single-station cross-component (SC) method to construct daily correlation functions, and dv/v was computed by the stretching method in the frequency band of 4–8 Hz. Moreover, we simulated dv/v using a near-surface layer of limited storage capacity and adjusted hydraulic conductivity (Ks). Our results indicate that peak dv/v at the downstream station was about four times that at the divide, with recovery to pre-typhoon levels taking 29 days, compared to 8 days at the divide station. Simulated Ks shows that hydraulic conductivity values higher than borehole-derived estimates are required to best capture the observed dv/v responses, indicating the preferential flow may be via colluvium and fractured bedrock in the study site. In short, a five-day typhoon produced nearly one month of NSGD, demonstrating that near-surface groundwater may function as a temporary storage zone for deep groundwater. These results demonstrate that ambient seismic noise can resolve short-term subsurface water dynamics during extreme events, offering new constraints on water residence times and aquifer structure that are relevant for disaster management and biogeochemical studies in mountainous watersheds.
How to cite: Tsai, C.-H., Yang, C., Illien, L., and Chen, L.-W.: Typhoon-driven Near-surface Groundwater Dynamics Revealed by Ambient Noise in the Mountain Watershed, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9607, https://doi.org/10.5194/egusphere-egu26-9607, 2026.
Fluvial seismology is an emerging field that exploits seismic signals generated by fluvial processes to monitor bedload transport and flow turbulence. Currently, most previous studies have focused on small mountain rivers, while seismic signatures from large rivers remain poorly explored. The Tiger Leaping Gorge is a deeply incised, narrow gorge in the upper Yangtze River characterized by extreme topographic relief and intense fluvial incision. Over a river length of approximately 20 km, the riverbed elevation drops by ~200 m, and the maximum discharge can reach ~5000 m³ s⁻¹, making the gorge an exceptional natural laboratory for investigating the coupling between seismic signals and hydrodynamic processes in large rivers. Until March 2022, we deployed 35 seismic stations along the riverbanks of the Tiger Leaping Gorge to continuously monitor the actively incising river segment.
We analyzed eight months of continuous seismic data along the river channel. In contrast to observations from small rivers, we identify two distinct and well-separated seismic energy bands at most stations. Temporal variations in both frequency bands show strong correlations with river discharge. We interpret the higher-frequency energy band as being primarily generated by small-scale eddy interacting with a rough riverbed, a process that appears to be particularly pronounced in large rivers. Building on existing models of turbulence-generated and bedload-generated seismic signals, we further tested different inversion approaches and applied them to all stations in the array. This allowed us to reconstruct the spatiotemporal variations in river stage and bedload transport across the study area.
Our results reveal that large rivers exhibit seismic signal characteristics controlled by distinct flow-related mechanisms, a phenomenon that has not been fully recognized in previous studies of small-scale rivers. Moreover, this study demonstrates that dense seismic arrays can resolve river dynamic processes at high spatial and temporal resolution, highlighting the potential of fluvial seismology for monitoring large river systems.
How to cite: Yan, X., Tang, H., Liu-Zeng, J., Turowski, J., Wang, Y., Yang, J., Wang, J., and Zhou, Q.: Seismic signatures of large river dynamics revealed by a dense seismic array at the Tiger Leaping Gorge of the Jinsha River, SE Tibet, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6957, https://doi.org/10.5194/egusphere-egu26-6957, 2026.
The planet’s surface is a restless orator: its fractures, failures and flows in response to climatic, tectonic and anthropogenic forcing inscribe narratives in seismic waveforms. Our limited success in decoding these narratives by classification and attribution of complex, incognito signals, threatens to leave Environmental Seismology data-rich but epistemologically impoverished. We propose that geomorphic signals can be formalized as language and tracked as evolving lineages. Hence, we develop a self-organizing, classification method with comprehension, resulting in an explainable, evolving phylogenetic tree. Unlike supervised classification methods that require exhaustive labels or clustering algorithms that conflate statistical similarity with physical kinship, our classification tool learns without labels, generalizes without forgetting physics, and explains without obfuscation.
We present a “glass-box” classifier, unsupervised in perception but supervised in its definition, that treats seismic data not as flat feature vectors but as structured, generative text, translating ground motion into a lexicon of geomorphic processes. Our system discovers its own alphabet, syntax, and semantics: autonomously constructing a taxonomic tree for seismic events while remaining interpretable. To do so, we break down the continuous seismic signal into discrete "phonemes." Multi-scale temporal descriptors, impulsive micro-textures (1.25 s), meso-scale envelope dynamics (5 s) and slow background trends (20 s), compose a multi-metric feature suite. These windows are fused into a tensor encoding nonlinear force interactions, then discretized through RVQ into context-aware symbols. Thus, we replace the geometric rigidity of static clustering with a dynamic evolutionary state space where signal classes behave as adapted species (rockfalls, landslides, debris flows, mine collapses, volcanic tremors, GLOFs, tectonic and glacial earthquakes, nuclear explosions, anthropogenic noise), governed by the Free Energy Principle. Similarity of process signals is tripartite: syntactic (grammar divergence), information-theoretic (surprisal), and algorithmic (NCD). This distinguishes events that look alike but have differing mechanisms: a debris flow and lahar may share "rumble" words, yet obey distinct physical grammars.
Modeled on Darwinian phylodynamics, the spectral species defined by their grammatical structure (causality), "metabolize" incoming data by minimizing thermodynamic surprise. The populations undergo sympatric speciation, hybridization, commensalism, and extinction, disentangling the "phylogenetic distance" (similarity) between superficially similar signals, and ultimately resulting in optimized classification. The algorithmic biomimicry of our approach outperforms static taxonomies that fail in non-stationary Earth systems without retraining.
Applied to a global, geologically heterogeneous inventory of >6000 curated records, preliminary results show phonemes reliability reaches 94–98% across stations and a >50% drop in articulation-structure complexity from noise to geomorphic events. The inferred phylogeny is physically meaningful, decoupling categories in distinct topological manifolds, allowing the classifier to reject false positives without supervision. Uncertainty is metabolized: high-aleatoric/low-epistemic signals (inherent noise) separate from low-aleatoric/high-epistemic anomalies (black swans: candidate new species). Free Energy scores, combine complexity and inaccuracy, outperform baselines in robustness to gaps, clipping, and dropout by over 20%.
The model is self-interpreting; rather than an opaque class label, it outputs a semantic sentence, exposing the decision path to a physically meaningful event classification, whilst also encapsulating information unique to specific events. Legible, self-improving classification turns detection into naming, and identity into process understanding.
How to cite: Ursica, S. and Hovius, N.: Seismic events classification using language syntax and biomimesis, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17578, https://doi.org/10.5194/egusphere-egu26-17578, 2026.
The Greenland Ice Sheet (GrIS) is a major contributor to global sea-level rise. However, predicting its future contribution remains complicated due to large uncertainties in modeling seasonal velocity variations. One of the key challenges is to better constrain the role of isolated subglacial water cavities. Over the melting season, water pressure within these cavities fluctuates, causing the glacier base to be coupled or decoupled from the bedrock. This process modulates basal friction, thereby influencing the velocity of the glacier.
Yet, the mechanisms by which these cavities depressurize by connecting to efficient drainage systems remain poorly understood, particularly for rapid drainage events where viscous creep cannot play a role (Mejia et al., 2021).
To investigate this phenomenon and identify key hydrological parameters, a dense seismic array (Gimbert et al., 2021; Nanni et al., 2021) and a GNSS station was deployed over purported subglacial lakes at Isunguata Sermia, West Greenland. In early September, the GNSS station highlighted a sharp decrease in ice surface elevation, accompanied by intense seismicity at the ice-bed interface. We used unsupervised machine learning to explore recorded seismic signals and identified Low-Frequency Icequakes (LFI), which do not follow classical rupture scaling laws. Additionally, many of these events were followed by a tremor with very low frequencies (~2 Hz). These tremors migrated spatially over time, following a diffusion pattern correlated with the ice surface subsidence, suggesting water migration at the ice-bed interface and thus, a lake drainage. However, the estimated diffusion coefficient is two orders of magnitude higher than predicted by Darcy's flow law. This suggests a hydrofracturing mechanism, facilitating rapid connections between multiple isolated cavities rather than a simple, localized lake drainage process.
Unlike previous interpretations of LFIs observed on glaciers (Thelen et al., 2013; Helmstetter, 2022), our findings suggest that these events do not represent stick-slip mechanisms at the glacier base but are instead, generated by fluid pressure diffusion due to the pressure difference between two isolated cavities. To further support this idea, we propose a theoretical forward model for seismic noise generation driven by pressure diffusion. The model is based on the geometric and hydrological characteristics of the cavities, including cavity width, inter-cavity spacing, the diffusion coefficient and the water volume. Accurate identification of these parameters, based on observations allows us to reproduce the key features of the observed power spectrum.
How to cite: Rousseau, H., Le Bot, J., Gimbert, F., Esfahani, R., Campillo, M., Doyle, S. H., Livingstone, S., Sole, A., Michel, A., Paris, N., and Le Bris, T.: Low frequency icequakes as the signature of transient subglacial water flow underneath the Greenland Ice Sheet, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18176, https://doi.org/10.5194/egusphere-egu26-18176, 2026.
Slope instabilities pose an increasing threat to populations and infrastructure across various regions worldwide. Therefore, a fundamental understanding of processes governing slope failure is critical for improving hazard mitigation. While remote-sensing and synthetic aperture radar methods effectively capture surface displacement, they provide limited information on subsurface dynamics. Seismic monitoring and imaging techniques can provide valuable complementary information on the internal structure, material properties, and time-dependent processes associated with unstable slopes.
We present a large-scale application of long-term Distributed Acoustic Sensing (DAS) measurements to investigate the spatial and temporal evolution of microseismicity at Cuolm da Vi (Central Switzerland), one of the largest active slope instabilities in the Alps. We deployed a 6.5 km long fibre-optic array to record continuous DAS data over a five-month period in spring 2023. Using a coherence-based detection method that exploits the dense spatial sampling of DAS, we identified 1,277 local seismic events. Event locations were obtained by adapting a matched field processing (MFP) approach to DAS observations, resulting in a comprehensive microseismic catalogue. The localisation workflow was validated through a controlled-source experiment and by comparison with a traveltime inversion of manually picked arrivals for selected events.
The resulting event distribution shows a pronounced spatial correspondence with known tectonic structures at Cuolm da Vi, particularly steeply dipping fracture systems, suggesting that much of the observed seismicity is linked to internal deformation processes related to a toppling movement. Clusters of elevated event density coincide with regions of reduced seismic velocities or strong velocity contrasts inferred from an independently derived three-dimensional velocity model. During the five-month observation time, the seismicity exhibits three distinct phases of elevated activity, with the first two closely following periods of intense precipitation and snowmelt. In addition, distinct spatial migration patterns of seismic activity emerge across different timescales.
The findings of our study demonstrate that DAS enables long-term monitoring of microseismic activity over spatially extensive and challenging Alpine terrain. The results provide new constraints on the internal structure and evolving dynamics of the Cuolm da Vi instability and additionally, highlight the potential of DAS-based seismic monitoring to improve hazard assessment and advance our understanding of deep-seated slope failure processes.
How to cite: Kiers, T., Schmelzbach, C., Grimm, J., Amann, F., Maurer, H., Edme, P., Bonanomi, Y., and Robertsson, J.: Fibre-Optic Monitoring of An Alpine Slope Instability Using Seismic Events: A Spatio-Temporal Analysis, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18693, https://doi.org/10.5194/egusphere-egu26-18693, 2026.
During the melt season, subglacial drainage systems typically evolve from a high-pressure, distributed system to a low-pressure, channelized network that progressively extends up-glacier from the terminus in response to meltwater availability. Resolving the spatial and temporal evolution of this transition remains challenging, particularly with small seismological networks, or even single stations.In spring 2023, we deployed three seismological stations along the central flow line of the southeast outlet glacier of the A. P. Olsen Ice Cap, northeast Greenland. The stations spanned the full vertical extent of the ablation zone and continuously recorded throughout the 2023 melt season.We apply and compare different seismological analysis techniques with the potential to detect changes in subglacial hydrological conditions. The seismic observations are interpreted in conjunction with meteorological data from two automatic weather stations located in the lower and upper ablation zone. We assess the capability of environmental seismological monitoring with single stations to track drainage system development in space and time.
How to cite: Binder, D., Mertl, S., Larsen, S. H., and Eibl, E. P. S.: Can we track the up-glacier migration of a subglacial channelized drainage system by means of environmental seismology?, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21131, https://doi.org/10.5194/egusphere-egu26-21131, 2026.
Understanding river sediment transport, and bedrock incision remains a major challenge in fluvial geomorphology Capturing their full temporal dynamics requires long-term monitoring of experimental catchments. This study explores the potential of Distributed Acoustic Sensing (DAS) technology to advance our understanding of fluvial sediment transport and riverbed evolution. DAS not only records strain or strain rates at meter-scale resolutions, similar to riverine dense geophone arrays but also captures a broad frequency range (mHz to kHz), comparable to hydrophones. Two experiments were conducted in meandering and straight artificial channels, with boundaries lined by waterproof membranes and stone protections, allowing for systematic effects of boundary and meandering shape. The experimental channels had a trapezoidal cross section, with widths ranging from approximately 2 to 4 m and bed slopes of 4–5°. During the experiments, the maximum flow depth reached about 0.3–0.4 m, the discharge ranged between 0.5 and 1 m³ s⁻¹, and the median grain size (D50) was approximately 10–20 mm. The experiments were monitored using a synchronized multi-sensor framework that combined UAV- and ground-based photogrammetry, particle tracking velocimetry, water-level gauges, stand-alone hydroacoustic sensor, riverine seismic dense array and DAS monitoring. Two fiber-optic burial configurations were examined for strain-rate sensing: (1) burial within a 30 cm thick sediment layer, and (2) installation beneath the armored riverbed (riprap) layer, allowing assessment of coupling conditions. Two-gauge lengths (2 m and 10 m) were also tested to evaluate their influence on strain-rate measurements. In our artificial channel experiments, the DAS measurements successfully captured high–spatiotemporal-resolution riverbed erosion and deposition dynamics. Fibers buried beneath the armored riverbed layer exhibited less sensitivity to riverbed morphological changes compared to those embedded within the sediment layer. In addition, the integration of DAS strain-rate, hydrophone, and riverbank seismic array data provided a comprehensive characterization of sediment transport processes across the channel. This study demonstrated that fluvial DAS enables continuous, high–spatiotemporal-resolution monitoring of sediment transport and riverbed evolution.
How to cite: Chao, W.-A., Hung, C.-Y., Chen, Y.-S., and Chen, S.-C.: Field-scale artificial channel experiments for fluvial DAS observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15481, https://doi.org/10.5194/egusphere-egu26-15481, 2026.
