AAAlice Abbate

Alice Abbate

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Alice Abbate has a total of 56 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Stochastic Emulation for Efficient Onshore Probabilistic Tsunami Hazard Assessment

    by Naveen Ragu Ramalingam, Alice Abbate, Erlend Briseid Storrøsten, Gareth Davies, Stefano Lorito, Manuela Volpe, Steven J. Gibbons, Fabrizio Romano, Finn Løvholt

    Machine learning is emerging as a promising strategy for modeling tsunami inundation at reduced computational cost. To enable probabilistic outputs that capture emulator uncertainty, we employ an ensemble of stochastic encoder‐decoder emulators with dropout‐based stochastic forward pass. Training and testing is conducted on an extensive Mediterranean Sea tsunami simulation data set comprising multiple earthquake source regions with variable magnitude, location, and focal mechanism, including near‐field events with local coseismic deformation. For the two target sites Catania and Siracusa in Sicily, we achieve good performance for relatively sparse training sets of only a few hundred events. We demonstrate the potential of our emulators to accurately model inundation from a range of tsunami sources needed to perform probabilistic tsunami hazard assessment (PTHA), reducing the computational cost of PTHA by a factor of 30–100, and estimating the uncertainty introduced by the emulator. Comparing probabilistic hazard maps and curves from our emulation with those from the full simulation ensemble and importance sampling Monte Carlo approach, we demonstrate the efficacy of emulators in capturing the spatial distribution and magnitude of inundation needed in overland tsunami hazard assessment.

  • Open Access

    From PTHA to Probability of Safe Evacuation: An Agent-Based Modelling (ABM) Use Case for EPOS Integrated Core Services - Distributed

    by Saeed Soltani, Fatemeh Jalayer, Stefano Lorito, Manuela Volpe, Julie Dugdale, Hossein Ebrahimian, Saman Ghaffarian, Alice Abbate

    This work introduces a methodology that links Probabilistic Tsunami Hazard Analysis (PTHA) with social-behavioral simulation to support risk-informed decision-making, where safe evacuation probabilities are evaluated by integrating hazard likelihood with agent-based estimates of evacuation success or failure. The proposed model integrates heterogeneous digital assets, including probabilistic hazard scenarios, exposure datasets and statistically derived behavioral parameters, within a unified workflow, with several inputs already available or foreseen through the EPOS data portal. The methodological core follows a Probabilistic Tsunami Risk Assessment (PTRA) formulation and is embedded within a Monte Carlo integration scheme, where evacuation metrics are derived from repeated simulations under random realizations of uncertain inputs.From a digital infrastructure perspective, this work explores the set-up and requirements for developing a prototype model as a use case for the EPOS Integrated Core Services-Distributed (ICS-D). Execution of PTRA using ABM requires access to distributed computing resources to support large scale and large numbers of agent-based simulations, server-side storage for hazard scenarios and exposure datasets, and statistical analysis/machine learning tools for the semi-automated transformation of raw information into meaningful behavioral inputs. Simulation outputs and geospatial products are then generated through Python and GIS-compatible visualization workflows.Looking ahead, such configuration offers a basis for a probabilistic tsunami evacuation modelling service capable of incorporating cascading multi-hazard effects, such as earthquake-induced damage that directly influences evacuation dynamics. By explicitly representing individuals, infrastructure, and their interactions within a shared environment, it enables iterative updating of time dependent conditions. This provides a natural pathway toward a Digital Twin–oriented framework for tsunami evacuation, supporting adaptive, decision-relevant risk analysis and align with the objectives of the Tsunami Thematic Core Service of EPOS.

  • Open Access

    Performance of Stochastic Tsunami Source Models Compared with Observations and Finite-Fault Inversions

    by Hafize Başak Bayraktar, Stefano Lorito, Antonio Scala, Gaetano Festa, Fabrizio Romano, Alice Abbate, Manuela Volpe, Thorne Lay, Carlos Sánchez-Linares, Patricio Catalan, et al.

    Various approaches exist to generate physically plausible tsunami source models in order to quantify source uncertainty in tsunami hazard assessments. Among these, stochastic slip models are widely used due to their low computational cost. Finite-fault models derived from inversion results are also sometimes used to anticipate the slip distributions of future ruptures.In this study, we compare synthetic tsunamis generated using a stochastic slip model with far-field DART observations from 14 tsunami events. We construct eight classes of source models based on combinations of two scaling relations (Murotani et al., 2013; Strasser et al., 2010), circular and rectangular rupture geometries, and depth-independent versus depth-dependent rigidity and coupling. For each class, we simulate a number of synthetic tsunamis using stochastic slip distributions of earthquakes of magnitude and location similar to those of the earthquakes that generated the 14 tsunamis. The results indicate that nearly all source model classes exhibit a mild—though not statistically significant—tendency to generate synthetic tsunamis which overestimate the observed tsunami amplitudes.We further conduct a quantitative comparison of slip distributions and tsunami time series from the best-fitting stochastic scenarios with those obtained from finite-fault teleseismic inversion models, some of which are constrained also by tsunami data. Overall, the best-fitting stochastic models reproduce observed tsunami waveforms more accurately than models derived from teleseismic-only inversions. However, for some events, specific slip patterns inferred from inversion models, such as an annular shape, cannot be adequately reproduced by the stochastic approach, leading to poorer fits to the observations.

  • Open Access

    Deployment of a Multi-Parameter Volcanic Tsunami Monitoring System Offshore Sciara del Fuoco, Stromboli

    by Antonio Costanza, Francesco Macaluso, Gioacchino Fertitta, Mauro Coltelli, Donifan Lazzaro, Marcello D'AGostino, Stefano Lorito, Alessandro Amato, Alessio Piatanesi, Sergio Bruni, et al.

    A fully integrated tsunami monitoring unit, funded by the Italian National Civil Protection Department, was installed in July 2025 by INGV, offshore the Sciara del Fuoco, Stromboli, under the coordination of two INGV centres, the CME (Centro Monitoraggio Eolie) and CAT (Centro Allerta Tsunami), and has been operating reliably since.The seafloor module, anchored to a reinforced-concrete deadweight, hosts multiple sensors, including a pressure sensor to detect rapid sea-level changes, a hydrophone to capture volcanic and underwater activity and an accelerometer to observe seafloor dynamics.Information is exchanged bidirectionally between the buoy and the seafloor module through an elastic electromechanical cable, also supplying energy to the seafloor module, and simultaneously providing elastic mooring. The elastic cable can extend up to 2.5 times its unstretched length while maintaining its mechanical and electrical integrity. To reach the installation depth, two cable segments were connected using an intermediate connection buoy. A mechanical release system can be triggered acoustically to release the seafloor module from the deadweight.The depth of the seafloor module, approximately 88 meters, was chosen so that pressure data can be converted into tsunami height values to limit as much as possible the deep-water high-wavenumber attenuation effects, as tsunamis induced by mass flows over the Sciara del Fuoco feature shorter periods signals than earthquake-generated tsunamis. Data streams are transmitted via radio and Wi-Fi to the shore station on Stromboli Island, then forwarded to the INGV-Osservatorio Etneo data center at the COA (Centro Operazioni Avanzate) in Stromboli. From there, the data are transmitted to the INGV-CAT. An STA/LTA algorithm, first introduced by Ripepe & Lacanna (2024) for the tsunami warning at Stromboli from the elastic beacons at a similar location, is applied experimentally to the data stream. The data will become in the future an important source of information for the tsunami warning system in Stromboli.

  • Open Access

    Exploring earthquake source uncertainty in probabilistic tsunami hazard assessment

    by Alice Abbate, Andrey Babeyko, Hafize Basak Bayraktar, Antonio Scala, Stefano Lorito, Nikos Kalligeris

    Tsunamis are among the most impactful natural hazards, yet their rarity results in incomplete historical and instrumental records. Tsunami hazard assessment is therefore strongly affected by uncertainties, mainly related to the source representation. For earthquake-generated tsunamis, the location of future ruptures and their rupture characteristics (geometry, kinematics, slip distribution) is poorly constrained, leading to some subjective choices regarding the source representation. A probabilistic approach allows us to formally incorporate these uncertainties and to calculate the probability that a given tsunami intensity measure will be exceeded at a target location within a specified time window.The MULTI-MAREX project established two living-labs in Greece, with the purpose of strengthening preparedness and awareness of natural hazards from marine environments. For these two sites, we estimate the offshore hazard from earthquake-generated tsunamis from different source representations. We adopt the regional probabilistic NEAMTHM18 model to select most representative sources based on de-aggregation analysis. These include interface subduction earthquakes, mainly associated with the Hellenic Arc, and both strike- and dip-slip crustal earthquakes distributed over the region. Source geometries are derived from the mean values of established scaling relationships between fault parameters and earthquake magnitude, and alternative scaling relationships. To explore the sensitivity of tsunami hazard estimates to earthquake source variability, we perturb the selected source geometries by considering further alternative scaling relationships and their associated uncertainties, rather than only the mean values.In addition, we provide a preliminary assessment of the impact of constraining scenarios to a mapped offshore fault in the EFSM20. This provides the basis to verify the effect of including more mapped faults in NEAMTHM18, which is an improvement in principle, provided that faults are well-mapped. This work complements ongoing research in Sicily (within a Transnational Access provided by the Geo-INQUIRE project), where the influence of source scaling laws on both offshore and onshore probabilistic tsunami hazard is explored using nested high-resolution grids. At the MULTI-MAREX sites, offshore-only analyses are performed, yet using much higher resolutions to simulate the offshore propagation.This work is contributing to enhancing project tsunami scenario databank by better accounting for source-related uncertaintiest, that finds applications in high-resolution inundation modelling for onshore tsunami hazard and virtual reality modelling.

  • Open Access

    Geohazards as serious gameplay: Immersive Virtual Environments from Real-World data Enable Story- and Game-Based Engagement with Modeled Marine Geohazard Scenarios.

    by Jan Oliver Eisermann, Felix Gross, Josephin Wolf, Alice Abbate, Andrey Babeyko, Christian Wagner-Ahlfs, Tom Kwasnitschka, Heidrun Kopp, Sebastian Krastel

    The MULTI-MAREX research mission, initiated by the German Marine Research Alliance (DAM), is establishing two living labs in Greece to study extreme marine geological events and related hazards. To address the challenges of communicating research outcomes and risk assessments, we have developed a workflow for creating virtual reconstructions of real study sites that transform complex geohazard scenarios into photorealistic immersive experiences. These virtual scenarios enhance situational awareness and facilitate meaningful and fact based engagement with experts, policymakers, and the public.Using a game engine as a real-time 3D rendering platform enables the integration of physics-based numerical simulations with real-world spatial data thus providing an immersive frontend to classical numerical models. Our focus is on developing workflows that support a semi-automated, asset-enhanced, immersive visualisation of geospatial data within this framework. These virtual environments synthesise numerical physical models with remote sensing data, including terrestrial and marine digital outcrop models derived from drone and submersible imagery, as well as hydroacoustic bathymetry. Digitally placed assets, such as high-resolution synthetic textures, vegetation, cars, urban furniture and buildings, enhance the visual appearance and help to bridge the gap between different data resolutions. Physics-based simulations of fluids, objects, collisions, destruction, lighting and weather further transform real-world data into photorealistic, interactive environments.By integrating numerical simulations via a custom data interface, we can visualise the effects of tsunamis, volcanic eruptions, extreme weather and wildfires with high fidelity. The framework used allows for a scalable approach across platforms, ranging from smartphones and desktop systems to head-mounted displays. These platforms ensure that visualisations and gameplay can be adapted to reach different stakeholders.Stakeholders can experience scenarios from multiple perspectives, such as first-person or external observer view, and freely explore the open-world virtual environment. Interactive storylines support learning by guiding stakeholders through the environment and different scenarios. Additionally, stakeholders can engage with task-based, competitive elements of serious gaming, such as starting in an everyday situation before a realistic scenario is triggered, and then identifying the fastest route to safety. Decisions can have consequences and can be reviewed at the end of the experience to assess choices and learn from mistakes, with virtual objects providing guidance throughout.Virtual environments are powerful tools for enhancing scientific analysis and stakeholder engagement, bridging the gap between complex geohazard science and effective stakeholder understanding. This supports informed decision-making and experience-based risk management.

  • Open Access

    Importance sampling of seismic tsunami sources with near-field emphasis for inundation PTHA: benchmarking with complete ensembles

    by Alice Abbate, Gareth Davies, Stefano Lorito, Nikos Kalligeris, Fabrizio Romano, Roberto Tonini, Manuela Volpe

    Site-specific Probabilistic Tsunami Hazard Assessment (PTHA) is a powerful tool for coastal planning against tsunami risk. However, its typically high computational demands led to the introduction of a Monte Carlo Stratified Importance Sampling (SIS) approach, which selects a representative subset of scenarios for numerical inundation simulations. We here empirically validate this sampling approach, for the first time to our knowledge, using an existing extensive dataset of numerical inundation simulations for two coastal sites in the Mediterranean Sea (Catania and Siracusa, both located in Sicily, Italy). Moreover, we propose a modified importance sampling function to prioritise seismic tsunami scenarios based on their arrival time at an offshore point near the target site, in addition to their wave amplitude and occurrence rate as leveraged in the previous work. This sampling function is applied separately in each earthquake magnitude bin, and allows denser sampling of near-field earthquakes to whose variations tsunamis are very sensitive.We compare the confidence intervals of the offshore PTHA estimates obtained with the new and the original importance sampling functions. Then, we benchmark our onshore PTHA estimates obtained with both functions against the inundation PTHA calculated using the full set of scenarios. We also test the assumption that onshore random errors follow a normal distribution, as found previously for the offshore case. As a result of the benchmarks, we find that the SIS approach works satisfactorily. Introducing the arrival time as an additional sampling factor enhances the precision of the estimates of both the mean and the percentiles for the two coastal sites considered. With this modification it is possible to deal efficiently with heterogeneous near-field earthquake sources involving coastal deformation at Catania and Siracusa, in addition to regional crustal and subduction sources. By comparing the sampling errors with the model (epistemic) uncertainty, an optimal trade-off between the number of simulations employed and the uncertainty of the PTHA model can be found, even for such a complex situation. A relatively small number of scenarios, on the order of a few thousand, is sufficient to perform site-specific PTHA for practical applications. These numbers correspond to 4-8\% of the already reduced ensembles used in previous assessments at the same sites.

  • Open Access

    Importance sampling of seismic tsunami sources with near-field emphasis for inundation PTHA: benchmarking with complete ensembles

    by A Abbate, G Davies, S Lorito, N Kalligeris, F Romano, R Tonini, M Volpe

    SUMMARY: Site-specific Probabilistic Tsunami Hazard Assessment (PTHA) is a powerful tool for coastal planning against tsunami risk. However, its typically high computational demands led to the introduction of a Monte Carlo Stratified Importance Sampling (SIS) approach, which selects a representative subset of scenarios for numerical inundation simulations. We here empirically validate this sampling approach, for the first time to our knowledge, using an existing extensive data set of numerical inundation simulations for two coastal sites in the Mediterranean Sea (Catania and Siracusa, both located in Sicily, Italy). Moreover, we propose a modified importance sampling function to prioritize seismic tsunami scenarios based on their arrival time at an offshore point near the target site, in addition to their wave amplitude and occurrence rate as leveraged in the previous work. This sampling function is applied separately in each earthquake magnitude bin, and allows denser sampling of near-field earthquakes to whose variations tsunamis are very sensitive. We compare the confidence intervals of the offshore PTHA estimates obtained with the new and the original importance sampling functions. Then, we benchmark our onshore PTHA estimates obtained with both functions against the inundation PTHA calculated using the full set of scenarios. We also test the assumption that onshore random errors follow a normal distribution, as found previously for the offshore case. As a result of the benchmarks, we find that the SIS approach works satisfactorily. Introducing the arrival time as an additional sampling factor enhances the precision of the estimates of both the mean and the percentiles for the two coastal sites considered. With this modification it is possible to deal efficiently with heterogeneous near-field earthquake sources involving coastal deformation at Catania and Siracusa, in addition to regional crustal and subduction sources. By comparing the sampling errors with the model (epistemic) uncertainty, an optimal trade-off between the number of simulations employed and the uncertainty of the PTHA model can be found, even for such a complex situation. A relatively small number of scenarios, on the order of a few thousand, is sufficient to perform site-specific PTHA for practical applications. These numbers correspond to 4–8 per cent of the already reduced ensembles used in previous assessments at the same sites.


  • Open Access

    Modeling optimal initial conditions for propagation of seismotectonic tsunamis: a database of smoothed unit sources based on an efficient and accurate numerical integration

    by Alice Abbate, José Manuel González Vida, Manuel J. Castro Díaz, Fabrizio Romano, Hafize Başak Bayraktar, Andrey Babeyko, Stefano Lorito

    The initial conditions for the numerical simulation of a seismically-induced tsunami are modeled by transferring the impulse produced by the co-seismic seafloor deformation to the sea-surface. In this process, the water column acts as a hydraulic filter for the smaller wavelengths. The numerical simulation of this process is computationally demanding; this makes the application of this filter unaffordable in studies that require a large number of simulations, such as the long-term probabilistic tsunami hazard assessment (PTHA). Here, we optimize the numerical modeling of the filter in the case of an instantaneous vertical seafloor deformation, given by an improper Fourier expansion integral in the wave number domain presented by Nosov and Kolesov (PAGEOPH, 2011); the contribution of elementary seafloor displacements can then be linearly combine to obtain a static tsunami initial condition. We first explore the convergence of the integral in one dimension, to identify the range of wavenumbers significantly contributing to the integral. We find that its support can be limited to , being H the sea-depth. We then compare several quadrature formulae, selecting the optimal one in terms of accuracy and efficiency. We grid the domain into cells of equal size and constant depth, and verify that the nonlinear effects are negligible when we recombine them to obtain the initial sea level displacement. In two dimensions, the integral is solved with the optimal quadrature and the results tested on the tsunamigenic Kuril doublet sequence - a megathrust and an outer-rise - occurred in the Central Kuril Islands in late 2006 to early 2007. We also consider the horizontal co-seismic deformation projected on a slope and a simple model of the inelastic deformation of the wedge, on a realistic bathymetry. The approach proposed results accurate and fast enough to be relevant for practical applications, taking a few seconds for solving a single cell, depending on the local depth, and ~3min to recombine ~91k elementary initial conditions. We finally build a database of elementary initial conditions, as a function of the local sea depth, which can be linearly combined to obtain a discretization of any sea floor displacement globally.

  • Open Access

    ML Surrogate for Tsunami Forecasting and Hazard Assessment in Eastern Sicily

    by Naveen Ragu Ramalingam, Erlend Briseid Storrøsten, Steven Gibbons, Kendra Johnson, Gareth Davies, Stefano Lorito, Alice Abbate, Manuela Volpe, Fabrizio Romano, Finn Løvholt, et al.

    Addressing the challenges associated with the high computational cost of tsunami inundation simulation has been a persistent issue, particularly in capturing earthquake source uncertainty and solving the nonlinear shallow water equations on high-resolution grids. This study aims to alleviate this computational burden by leveraging machine learning surrogates. Further, evaluating these ML models is often hindered by their black-box nature and the limited size of training and testing datasets, posing challenges for practitioners. We propose an encoder-decoder neural network where offshore tsunami waveforms and local co-seismic deformation fields serve as the basis for predicting high-resolution inundation maps at 10m grids. The model is applied to the coastal region of Catania in Sicily, Italy, integrating diverse earthquake scenarios from a large simulation dataset of 53,550 tsunamigenic events in the Mediterranean Sea. We adopt a pretraining-fine-tuning approach for building the machine learning surrogate and address crucial questions regarding the efficient selection of training scenarios, model design, and training. Leveraging this large simulation dataset, we identify specific locations, scenarios and model conditions where the machine-learning surrogate demonstrates sufficient accuracy and reliability. This provides an efficient mechanism for long-term tsunami hazard assessment or urgent tsunami prediction in real-time situations.