The interdisciplinary study of volcanic processes, which extend across all timescales and lengths, requires a multitude of approaches, ranging from analogue and numerical modelling to observations and fieldwork and extending to mathematics. A conference was held at the East African Institute for Fundamental Research, affiliated with the University of Rwanda, a country which, along with the Democratic Republic of Congo, presents a unique geodynamic context. Located along the East African Rift, an active seismic region, Rwanda is close to two of Africa's most active volcanoes, including Nyiragongo, which overlooks Lake Kivu, a deep volcanic lake rich in dissolved carbon dioxide and methane, the latter of which is used for electricity generation. In this context, the conference addressed many “classic” volcanological topics and their modern advances, such as multiphase lava flows, subsurface magma propagation, seismic and deformation signals from a volcano, modelling of volcanic emission dispersion, and volcanic lakes. Yet, it broadened the discussion to the volcanic particle-water interface and its impact on soils, volcanoes and climate change, and volcanoes and health. This article aims to highlight and share the richness of the integration and interconnectedness of the various questions related to a volcanic environment, as well as their impact on society. Ultimately, this conference also demonstrated the importance of promoting science in Africa and developing countries so that the next generation of African researchers is equipped to address the challenges facing their nations.

Abstract The assessment of volcanic hazards is crucial to develop effective emergency plans, especially for volcanoes close to urban areas or under air traffic routes. Impact assessment for expected scenarios relies on underlying numerical models that require eruption source parameters as inputs, and forecasts drastically depend on their robust reconstruction during ongoing events. We apply a novel tephra deposit inversion workflow built on ensemble methods and data assimilation techniques to reconstruct the explosive events that occurred at Mt Etna, Italy, between 3 and 5 December 2015. Based on results from previous studies, we reconstruct this eruption using the Gaussian with Non‐negative Constraints data assimilation method. Results agree well with independent observations and highlight the potential for automatized procedures in volcanic hazard assessment.

Volcanic clouds threaten aviation and international collaborative efforts for monitoring and forecasting their dispersion have been in place for almost 30 years under the framework of the International Airways Volcano Watch (IAVW) Program. The 9 Volcanic Ash Advisory Centers (VAACs) play an important role in producing advisories for international civil aviation stakeholders on the presence and forecast of volcanic ash dispersion and transport. The Buenos Aires VAAC, which operates within the Argentinean National Meteorological Service under a regional air navigation agreement, has substantially improved its numerical modelling strategy in recent years. In parallel, the increasing user’s demand to navigate in low-contaminated areas based on impact evidence have pushed towards the development of Quantitative Volcanic Ash Information (QVA), which VAACs are starting to provide operationally during 2026 for significant eruptions. QVA consists of new deterministic and probabilistic forecast products based on ash concentration thresholds impacting aircraft operations and air navigation. The Buenos Aires VAAC has developed its first QVA system and this manuscript presents recent progress and its application on a case study. By the end of this decade, it is expected to extend the scope of QVA to all volcanic clouds, increasing its resolution and deprecating the current Volcanic Ash Advisories. Finally, we discuss the future work needed to achieve this objective.

Volcanic activity can pose a serious threat to nearby populations, as continuous gas emissions remain dangerous even in the absence of eruptions. Nyiragongo and Nyamuragira volcanoes located in the East African Rift are among the largest global emitters of SO2. Given the various environmental, climatic and health impacts of SO2, studying its dispersion is important. In our study we use FALL3D model (Folch et al., 2020), an Eulerian atmospheric dispersal model that solves the advection-diffusion-sedimentation equation, combined with ensemble-based data assimilation technique to reduce uncertainties in eruption source parameters to simulate SO2 dispersion during both eruptive and passive degassing phases.Satellite observations from TROPOMI are processed using a trained AI algorithm based on machine learning that automatically detects and quantifies volcanic SO2 emissions in near real-time filtering out non-volcanic sources (Corradino et al., 2024). The meteorological data used are from ERA5 reanalysis dataset.Literature studies (e.g.Mingari et al., 2022) show that the inclusion of the satellite data in the model greatly improves the dispersion forecasts. Building on these results we aim to improve the dispersion forecasts of SO2 from Nyiragongo volcano and develop probabilistic hazard maps of SO2 exposure enabling an uncertainty informed assessment of potential impacts on populations and infrastructure surrounding the volcanoes. Our study will demonstrate the potential of combining observational data, numerical modeling, and ensemble-based data assimilation to improve volcanic hazard monitoring.

Digital twin prototypes now exist for assessing and aiding disaster response to geophysical hazards, such as volcanic eruptions, tsunamis, and earthquakes. We propose a systematic assessment framework to help determine the accuracy, transparency, reproducibility, and accountability of machine learning components within these emerging prototypes.

Short and long-term probabilistic volcano hazard assessments entail the fusion of data from multiple sources with the realisation and subsequent combination of hundreds/thousands of scenarios spanning the range of system uncertainties (model inputs and parameterisations, boundary conditions, etc). Computational workflows are middleware software layers that manage and orchestrate in an automated way the multiple steps and tasks involved in this process, from data acquisition and preparation, to model executions and post-process in centralised (HPC) and/or cloud computing infrastructures. This contribution presents some examples from on-going European projects tackling computational geohazards on HPC/cloud infrastructures. For the short-term hazard assessment, the DT-GEO project (2022-2025, Grant Agreement No 101058129) is implementing a number of workflows conducting precise data-informed early warning systems and hazard assessments by harnessing world-class computational (FENIX, EuroHPC) and data (EPOS) research infrastructures. The volcano-related workflows in DT-GEO include: (i) merging of multi-parametric data from ground-based and remote observation systems (on-site monitoring networks and satellites) with global modelling of magma and rock dynamics and with AI approach; (ii) merging of real-time geostationary satellite observations with the FALL3D model to generate deterministic and ensemble-based probabilistic forecast products; (iii) merging of real-time multi-parametric data from ground-based and remote observation systems with deterministic modelling of lava flow propagation and inundation areas and; (iv) air-quality data and AI in a volcanic gas dispersal forecast context to improve operational Early Warning Systems. On the other hand, the EuroHPC ChEESE Center of Excellence (CoE) is conducting an ensemble-based volcanic dispersal across multiple scales that will lead to the first European tephra hazard map at scale covering, simultaneously, long-range dispersal and short-range fallout telescopically. This will be integrated in the EPOS Volcanic Observations TCS (VO-TCS). All these initiatives liaise, align, and synergise with EPOS and longer-term mission-like initiatives like Destination Earth.

Abstract. Ensemble-based modelling of the atmospheric dispersal of volcanic clouds enables more realistic forecasts by explicitly accounting for uncertainties in eruption source parameters, meteorological data, and systematic errors in transport models. Many ensemble applications, including quantification of forecast uncertainties, data assimilation, or probabilistic hazard assessments, require a large number of members to mitigate sampling errors and to properly capture probability distributions. However, running large ensembles with Volcanic Ash Transport and Dispersal (VATD) models can be computationally demanding, even for high-performance computing clusters. As a result, operational forecasting is typically restricted to smaller ensembles in order to fit time-to-solution requirements. In contrast, generative AI models can produce large volumes of physically-consistent data samples with minimal computational cost. In this work, a convolutional Variational AutoEncoder (VAE) is trained on an ensemble of 256 forecasts simulated with the FALL3D model and subsequently used to generate larger ensembles, effectively augmenting physics-based ensemble modelling capacity. Ensembles with up to 8192 members were generated nearly instantaneously using the trained neural network, with no reliance on HPC resources. The statistical properties of the expanded ensembles are characterised in detail, and the VAE performance is evaluated against a test dataset composed of 2048 numerical simulations. The VAE-generated ensembles closely approximate the actual (target) probability distribution as well as key sample statistics, such as ensemble mean and spread, with minimal degradation in the evaluation metrics. Finally, we discuss possible future applications of this work, including latent space data assimilation via deep learning.

Volcanic gas emission represents a source of hazard to humans and the environment. They occur both during volcanic unrest, eruptions and in quiescent stages of the volcanic activity. Therefore, it is a widespread and frequent threat. Many gas species (e.g. CO2, SO2) can affect human health and even threaten life at concentrations and doses above species-specific thresholds. Depending on the relative buoyancy at the emission location, volcanic gas emissions can be generally classified as dilute passive degassing and dense gas flow. Numerical simulations of gas dispersion involve a workflow that can be complex and time-consuming, since it starts with the modelling of the wind field, proceeds with the gas dispersion simulation and ends with the postprocessing stage. This process should be repeated several times (hundreds to thousands) for probabilistic volcanic hazard applications, in which the uncertainty of the relevant input parameters (e.g. wind field, emission rates and source locations) is explored to obtain probabilistic outputs. Here we present VIGIL, a Python simulation tool that manages the gas dispersion simulation workflow and is interfaced with two dispersion models: a dilute (DISGAS) and a dense gas (TWODEE-2) dispersion model. We show results from different applications showcasing the various capabilities of VIGIL.

The emission of volcanic gases can occur during both eruptive and quiescent stages of volcanic activity, affecting air quality in the surrounding areas and threatening human health when the concentrations exceed species-specific thresholds. In this regard, quantitative studies of model validation are essential before applying a simulator for probabilistic volcanic hazard assessment. Here, we provide a model validation aimed at testing the accuracy in providing realistic values of CO2 concentration at two active volcanic sites affected by persistent passive gas dispersion: La Solfatara (a maar crater within Campi Flegrei caldera, Italy) and Caldeiras da Ribeira Grande located in the north flank of Fogo volcano (São Miguel Island, Azores). We used published and original CO2 flux data as input for numerical simulations run through VIGIL, an open-source workflow for parallel simulations and probabilistic output using two Eulerian models, which account for the passive and gravity-driven gas transport, respectively. At Solfatara, we compared a 1-month-long simulation during June 2020 with CO2 concentration acquired by the INGV measurement station at 4 m from the ground in a selected point close to Pisciarelli vent: Our results showed a good correlation between the daily simulated and observed averages of CO2 concentrations. At Caldeiras da Ribeira Grande, we quantified the CO2 concentration at 43 tracking points, each referring to a specific acquisition (in space and time) during 13 selected days in July 2021. The comparison between the 1-month-long simulation and the observed data provided acceptable accordance. In both cases, we noted that the daily averaged concentrations provided by the model do not exceed the gas hazardous threshold limits. However, for shorter timescales (hours), a higher data acquisition rate is needed for future investigation.

The second phase (2023-2026) of the EuroHPC Center of Excellence for Exascale in Solid Earth (ChEESE-2P), funded by HORIZON-EUROHPC-JU-2021-COE-01 under the Grant Agreement No 101093038, will prepare 11 European flagship codes from different geoscience domains. Codes will be optimised in terms of performance on different types of accelerators, scalability, containerisation, and continuous deployment and portability across tier-0/tier-1 European systems as well as on novel hardware architectures emerging from the EuroHPC Pilots (EuPEX/OpenSequana and EuPilot/RISC-V) by co-designing with mini-apps. Flagship codes and workflows will be combined to farm a new generation of 9 Pilot Demonstrators (PDs) and 15 related Simulation Cases (SCs) representing capability and capacity computational challenges selected based on their scientific importance, social relevance, or urgency. On the other hand, the first phase of ChEESE was pivotal in leveraging an ecosystem of European projects and initiatives tackling computational geohazards that will benefit from current and upcoming exascale EuroHPC infrastructures. In particular, Geo-INQUIRE (2022-2024, GA No 101058518) and DT-GEO (2022-2025, GA No 101058129) are two on-going Horizon Europe projects relevant to the Solid Earth ecosystem. The former will provide virtual and trans-national service access to data and state-of-the-art numerical models and workflows for monitoring and simulation of the dynamic processes in the geosphere at unprecedented levels of detail and precision. The later will deploy a prototype Digital Twin (DT) on geophysical extremes including 12 self-contained Digital Twin Components (DTCs) addressing specific hazardous phenomena from volcanoes, tsunamis, earthquakes, and anthropogenically-induced extremes to conduct precise data-informed early warning systems, forecasts, and hazard assessments across multiple time scales. All these initiatives liaise, align, and synergise with EPOS and longer-term mission-like initiatives like Destination Earth.

The Millennium Eruption of Paektu volcano, on the border of China and North Korea, generated tephra deposits that extend >1000 km from the vent, making it one of the largest eruptions in historical times. Based on observed thicknesses and compositions of the deposits, the widespread tephra dispersal is attributed to two eruption phases fuelled by chemically distinct magmas that produced both pyroclastic flows and fallout deposits. We used an ensemble-based method with a dual step inversion, in combination with the FALL3D atmospheric tephra transport model, to constrain these two different phases. The volume of the two distinct phases has been calculated. The results indicate that about 3-16 km 3 (with a best estimate of 7.2 km 3 ) and 4-20 km 3 (with a best estimate of 9.3 km 3 ) of magma were erupted during the comendite and trachyte phases of the eruption, respectively. Eruption rates of up to 4 × 10 8 kg/s generated plumes that extended 30-40 km up into the stratosphere during each phase.

A Digital Twin Component (DTC) provides users with digital replicas of different components of the Earth system through unified frameworks integrating real-time observations and state-of-the-art numerical models. Scenarios of extreme events for natural hazards can be studied from the genesis to propagation and impacts using a single DTC or multiple coupled DTCs. The EU DT-GEO project (2022-2025) is implementing a prototype digital twin on geophysical extremes consisting of 12 interrelated Digital Twin Components, intended as self-contained and containerised software entities embedding numerical model codes, management of real-time data streams and data assimilation methodologies. DTCs can be deployed and executed in centralized High Performance Computing (HPC) and cloud computing Research Infrastructures (RIs). In particular, the DTC-V2 is implementing an ensemble-based automated operational system for deterministic and probabilistic forecast of long-range ash dispersal and local-scale tephra fallout. The system continuously screens different ground-based and satellite-based data sources and a workflow is automatically triggered by a volcanic eruption to stream and pre-process data, its ingestion into the FALL3D dispersal model, a centralized or distributed HPC model execution, and the post-processing step. The DTCs will provide capability for analyses, forecasts, uncertainty quantification, and "what if" scenarios for natural and anthropogenic hazards, with a long-term ambition towards the Destination Earth mission-like initiative.

Explosive volcanic eruptions inject hot mixtures of solid particles (tephra) and gasses into the atmosphere. Entraining ambient air, these mixtures can form plumes rising tens of kilometers until they spread laterally, forming umbrella clouds. While the largest clasts tend to settle in proximity to the volcano, the smallest fragments, commonly referred to as ash (≤2 mm in diameter), can be transported over long distances, forming volcanic clouds. Tephra plumes and clouds pose significant hazards to human society, affecting infrastructure, and human health through deposition on the ground or airborne suspension at low altitudes. Additionally, volcanic clouds are a threat to aviation, during both high-risk actions such as take-off and landing and at standard cruising altitudes. The ability to monitor and forecast tephra plumes and clouds is fundamental to mitigate the hazard associated with explosive eruptions. To that end, various monitoring techniques, ranging from ground-based instruments to sensors on-board satellites, and forecasting strategies, based on running numerical models to track the position of volcanic clouds, are efficiently employed. However, some limitations still exist, mainly due to the high unpredictability and variability of explosive eruptions, as well as the multiphase and complex nature of volcanic plumes. In the next decades, advances in monitoring and computational capabilities are expected to address these limitations and significantly improve the mitigation of the risk associated with tephra plumes and clouds.

The atmospheric dispersion of gases (of natural or industrial origins) can be very hazardous to life and the environment if the concentration of some gas species overcome specie-specific thresholds. In this context, the natural variability associated to the natural phenomena has to be explored to provide robust probabilistic gas dispersion hazard assessments.VIGIL-1.3 (automatic probabilistic VolcanIc Gas dIspersion modeLling) is a Python simulation tool born to automatize the complex and time-consuming simulation workflow required to process a large number of gas dispersion numerical simulations. It is interfaced with two models: a dilute (DISGAS) and a dense gas (TWODEE-2) dispersion model. The former is used when the density of the gas plume at the source is lower than the atmospheric density (e.g. fumaroles), the latter when the gas density is higher than the atmosphere and the gas accumulates on the ground and may flow due to the density contrast with the atmosphere to form a gravity current (e.g. cold CO2 flows).In the enhancement of the code towards a higher-scale computing, here we present the ongoing improvements aimed to extend some code functionalities such as memory management, modularity revision, and full-ensemble uncertainty on gas dispersal scenarios (e.g. sampling techniques for gas fluxes and source locations).Optimizations are also provided in terms of tracking errors, redesignation of the input file, validation of data provided by the users, and addition of the Latin hypercube sampling (LHS) for the post-processing of model outputs.All these new features will be issued in the future release of the code (VIGIL-2.0) in order to facilitate the users which could run VIGIL on laptops or large supercomputer, and to widen the spectrum of model applications from routinely operational forecast of volcanic gas to long-term hazard and/or risk assessments purposes.

Destination Earth initiative pursues the implementation of a digital model of the Earth. With the aim to help understand and simulate the evolution and behavior of the Earth system components, to aid in better forecasting the impacts on human system processes, ecosystem processes and their interaction. The current state of the art technologies in numerical computations (HPC), data infrastructures (involving data storage, data access, data analysis), enable the possibility of developing numerical clones mimicking Earth’s geophysical extreme phenomena.A Digital Twin for GEOphysical extremes (DT-GEO),is a new EU project funded under the Horizon Europe programme (2022-2025), with the objective of developing a prototype for a digital twin on geophysical extremes including earthquakes, volcanoes, tsunamis, and anthropogenic-induced extreme events. It will enable analyses, forecasts, and responses to “what if” scenarios for natural hazards from their genesis phases and across their temporal and spatial scales. The project consortium brings together world-class computational and data Research Infrastructures (RIs), operational monitoring networks, and leading-edge research and academia partnerships in various fields of geophysics. It mergesthe latest outcomes from other European projects and, Centers of Excellence. DT-GEO will deploy and test 12 Digital Twin Components (DTCs). These will be self-contained entities embedding flagship simulation codes, Artificial Intelligence layers, large volumes of (real-time) data streams from and into data-lakes, data assimilation methodologies, and overarching workflows for deployment and execution of single or coupled DTCs in centralized HPC and virtual cloud computing Ris. (DT-GEO: A Digital Twin for GEOphysical extremes, project ID 101058129)

: Operational forecasts of volcanic clouds are a key decision-making component for civil protection agencies and aviation authorities during the occurrence of volcanic crises. Quantitative operational forecasts are challenging due to the large uncertainties that typically exist on characterising volcanic emissions in real time. Data assimilation, including source term inversion, has long been recognised by the scientific community as a mechanism to reduce quantitative forecast errors. In terms of research, substantial progress has occurred during the last decade following the recommendations from the ash dispersal forecast workshops organised by the International Union of Geodesy and Geophysics (IUGG) and the World Meteorological Organization (WMO). The meetings held in Geneva in 2010–11 in the aftermath of the 2010 Eyjafjallajökull eruption identified data assimilation as a research priority. This Chapter reviews the scientific progress and its transfer into operations, which is leveraging a new generation of operational forecast products.

<p>DT-GEO is a project proposed to deal with natural or anthropogenically induced geohazards (earthquakes, volcanoes, landslides and tsunamis) by deploying a Digital Twin of the planet. The prototype will provide a way to visualize, manipulate and understand the response to hypothetical or on-going events by integrating data acquisition  and models. </p> <p>Due to the complexity of the development, the project has been divided into different work packages and components. The volcanic phenomena package includes 4 Digital Twin Components (DTCs): volcanic unrest, volcanic ash clouds and ground accumulations, lava flows, and volcanic gas dispersal. The volcanic ash and dispersal deposition component implements a workflow for atmospheric dispersal and ground deposition forecast systems. The workflow is composed of four general units. The first one is the Numerical Weather Prediction (NWP) acquisition (provided by external institutions) refers to both:  automatic obtention of the forecast (up to few days ahead) or the reanalysis (preprocess data from the past) in global or regional scales at different resolutions. Then, the Triggering and Eruption Source Parameters (ESP) is based on predefined communications channels and prioritized by an accuracy rank. The FALL3D model setup and run ensemble simulations, resulting from perturbing ESP values within a range. Finally, the postprocess refers to the compilation of the simulations into hazard maps.</p>

The second phase (2023-2026) of the Center of Excellence for Exascale in Solid Earth (ChEESE-2P), funded by HORIZON-EUROHPC-JU-2021-COE-01 under the Grant Agreement No 101093038, will prepare 11 European flagship codes from different geoscience domains (computational seismology, magnetohydrodynamics, physical volcanology, tsunamis, geodynamics, and glacier hazards). Codes will be optimised in terms of performance on different types of accelerators, scalability, containerisation, and continuous deployment and portability across tier-0/tier-1 European systems as well as on novel hardware architectures emerging from the EuroHPC Pilots (EuPEX/OpenSequana and EuPilot/RISC-V) by co-designing with mini-apps. Flagship codes and workflows will be combined to farm a new generation of 9 Pilot Demonstrators (PDs) and 15 related Simulation Cases (SCs) representing capability and capacity computational challenges selected based on their scientific importance, social relevance, or urgency. The SCs will produce relevant EOSC-enabled datasets and enable services on aspects of geohazards like urgent computing, early warning forecast, hazard assessment, or fostering an emergency access mode in EuroHPC systems for geohazardous events including access policy recommendations. Finally, ChEESE-2P will liaise, align, and synergise with other domain-specific European projects on digital twins and longer-term mission-like initiatives like Destination Earth.

In recent years, there has been a growing interest in ensemble approaches for modelling the atmospheric transport of volcanic aerosol, ash, and lapilli (tephra). The development of such techniques enables the exploration of novel methods for incorporating real observations into tephra dispersal models. However, traditional data assimilation algorithms, including ensemble Kalman filter (EnKF) methods, can yield suboptimal state estimates for positive-definite variables such as those related to volcanic aerosols and tephra deposits. This study proposes two new ensemble-based data assimilation techniques for semi-positive-definite variables with highly skewed uncertainty distributions, including aerosol concentrations and tephra deposit mass loading: the Gaussian with non-negative constraints (GNC) and gamma inverse-gamma (GIG) methods. The proposed methods are applied to reconstruct the tephra fallout deposit resulting from the 2015 Calbuco eruption using an ensemble of 256 runs performed with the FALL3D dispersal model. An assessment of the methodologies is conducted considering two independent datasets of deposit thickness measurements: an assimilation dataset and a validation dataset. Different evaluation metrics (e.g. RMSE, MBE, and SMAPE) are computed for the validation dataset, and the results are compared to two references: the ensemble prior mean and the EnKF analysis. Results show that the assimilation leads to a significant improvement over the first-guess results obtained from the simple ensemble forecast. The evidence from this study suggests that the GNC method was the most skilful approach and represents a promising alternative for assimilation of volcanic fallout data. The spatial distributions of the tephra fallout deposit thickness and volume according to the GNC analysis are in good agreement with estimations based on field measurements and isopach maps reported in previous studies. On the other hand, although it is an interesting approach, the GIG method failed to improve the EnKF analysis.

The EU Center of Excellence for Exascale in Solid Earth (ChEESE) develops exascale transition capabilities in the domain of Solid Earth, an area of geophysics rich in computational challenges embracing different approaches to exascale (capability, capacity, and urgent computing). The first implementation phase of the project (ChEESE-1P; 2018–2022) addressed scientific and technical computational challenges in seismology, tsunami science, volcanology, and magnetohydrodynamics, in order to understand the phenomena, anticipate the impact of natural disasters, and contribute to risk management. The project initiated the optimisation of 10 community flagship codes for the upcoming exascale systems and implemented 12 Pilot Demonstrators that combine the flagship codes with dedicated workflows in order to address the underlying capability and capacity computational challenges. Pilot Demonstrators reaching more mature Technology Readiness Levels (TRLs) were further enabled in operational service environments on critical aspects of geohazards such as long-term and short-term probabilistic hazard assessment, urgent computing, and early warning and probabilistic forecasting. Partnership and service co-design with members of the project Industry and User Board (IUB) leveraged the uptake of results across multiple research institutions, academia, industry, and public governance bodies (e.g. civil protection agencies). This article summarises the implementation strategy and the results from ChEESE-1P, outlining also the underpinning concepts and the roadmap for the on-going second project implementation phase (ChEESE-2P; 2023–2026).