Volcanic Hazard Assessment Research Articles

The recent volcanism of Flores Island (Azores): Stratigraphy and eruptive history of Funda Volcanic System

Monogenetic volcanoes occur in many different geotectonic settings and are usually small and short-lived. They can experience a variety of eruptive styles, even during the same eruption. In monogenetic volcanic fields, volcanism usually migrates to different locations over time, making volcanic hazard assessment very challenging. The eruptive history of a volcanic region, including the size, style, and location of previous eruptions, provides valuable information to help predict the behaviour of future volcanic events and their associated hazards. Here, we reconstruct for the first time the eruptive history of the Funda Volcanic System (FVS), one of the most recent (∼3 ka) monogenetic eruptive centres of Flores Island (Azores), based on a detailed tephrostratigraphic work coupled with geochemical analysis of glass shards and radiocarbon dating. We identified at least three volcanic events at FVS spaced by time intervals of ∼100 yr. The first event (3430 cal yr BP) was a small Strombolian eruption, the second event (3330 cal yr BP) started as a violent Strombolian eruption and may have ended as phreatomagmatic, and the third event (3250 cal yr BP) was exclusively phreatomagmatic. Our results demonstrate that volcanism at the FVS was more prolonged and recurrent than previously reported. Moreover, we show that the FVS experienced different eruptive styles in a short timeframe, ranging from small basaltic eruptions to violent explosive phreatomagmatic events. Such diversity of eruptive styles results in different volcanic products, which have different hazard implications. Our new results contribute to the knowledge of the recent volcanic activity of Flores Island, and we anticipate them to be of paramount importance for future volcanic hazard assessment studies.

On the feasibility and usefulness of high-performance computing in probabilistic volcanic hazard assessment: An application to tephra hazard from Campi Flegrei

For active volcanoes, knowledge about probabilities of eruption and impacted areas becomes valuable information for decision-makers to develop short- and long-term emergency plans, for which probabilistic volcanic hazard assessment (PVHA) is needed. High-resolution or spatially extended PVHA requires extreme-scale high-performance computing systems. Within the framework of ChEESE (Center of Excellence for Exascale in Solid Earth; www.cheese-coe.eu), an effort was made to generate exascale-suitable codes and workflows to collect and process in some hours the large amount of data that a quality PVHA requires. To this end, we created an optimized HPC-based workflow coined PVHA_HPC-WF to develop PVHA for a volcano. This tool uses the Bayesian event tree methodology to calculate eruption probabilities, vent-opening location(s), and eruptive source parameters (ESPs) based on volcano history, monitoring system data, and meteorological conditions. Then, the tool interacts with the chosen hazard model, performing a simulation for each ESP set or volcanic scenario (VS). Finally, the resulting information is processed by proof-of-concept-subjected high-performance data analytics (HPDA) scripts, producing the hazard maps which describe the probability over time of exceeding critical thresholds at each location in the investigated geographical domain. Although PVHA_HPC-WF can be adapted to other hazards, we focus here on tephra (i.e., lapilli and ash) transport and deposition. As an application, we performed PVHA for Campi Flegrei (CF), Italy, an active volcano located in one of the most densely inhabited areas in Europe and under busy air traffic routes. CF is currently in unrest, classified as being in an attention level by the Italian Civil Protection. We consider an approximate 2,000 × 2,000 × 40 km computational domain with 2 km grid resolution in the horizontal and 40 vertical levels, centered in CF. To explore the natural variability and uncertainty of the eruptive conditions, we consider a large number of VSs allowing us to include those of low probability but high impact, and simulations of tephra dispersal are performed for each of them using the FALL3D model. Results show the potential of HPC to timely execute a vast range of simulations of complex numerical models in large high-resolution computational domains and analyze great volumes of data to obtain quality hazard maps.

How long-term hazard assessment may help to anticipate volcanic eruptions: The case of La Palma eruption 2021 (Canary Islands)

Long-term hazard assessment is an essential action that should be undertaken at all active volcanoes and/or volcanic areas as an indispensable contribution to reduce volcanic risk. It allows anticipating new eruptions responding to two of the main questions of volcano forecasting: where and how the next eruption will be. In the case of the Canary Islands, an active volcanic region that receives millions of visitors every year, and which has been recently severely impacted by a small size eruption on the island of La Palma in 2021, a systematic long-term hazard assessment is still pending, despite this being foreseen in its management plan to face volcanic threats. An essential basic long-term volcanic hazard assessment, conducted six years before this last eruption, was able to anticipate the location of the new vents and the eruptive style, including the extent of its main potential hazards. This long-term research was developed in the frame of a Master thesis, which was aimed at justifying the viability of a particular systematic long-term volcanic assessment methodology to be applied to the Canary Islands and elsewhere. Despite this particular study was not used to manage the La Palma 2021 eruption, it demonstrates how beneficial such studies may be to forecast new eruptions and to reduce their associated risk, and represents a new contribution to develop adequate territorial and management planning where we provide a qualitative hazard map for the Island of La Palma.

Stratigraphy and eruption history of maars in the Clear Lake Volcanic Field, California

The Clear Lake Volcanic Field (CLVF) is the northernmost and youngest field in a chain of volcanic fields in the California Coast Range mountains. Effusive and explosive volcanic activity in the field has spanned at least 2.1 million years, with the youngest eruptions comprising a series of maar craters at the edges of, and within, Clear Lake itself. This work documents the first direct ages for many of these maar deposits and builds the stratigraphic basis for interpreting eruptive processes and dynamics of the young eruptions which produced them. Detailed stratigraphy has distinguished maar eruption products from pyroclastic deposits (monolithologic falls and flows, previously mapped together with maars as a single unit) and established a set of six eruption facies from maar deposit lithology, grain size parameters, and depositional structures. Radiocarbon dates from carbon films found on clasts at three outcrops have constrained several of these maar eruptions to ∼8,500–13,500 years BP, coinciding with eruptive periods previously estimated based on lake core tephrachronology. Part of this period also coincides with indigenous inhabitation (<12,000 years BP), which suggests that oral histories of Pomo and other local tribes may contain descriptions of volcanic phenomena experienced by local residents of the CLVF. Collaboration between volcanologists and indigenous historians may add a valuable human dimension to the youngest eruptions of the Clear Lake Volcanic Field, and help inform future volcanic hazard assessment.

Magma pressurization sustains ongoing eruptive episode at dome-building Soufrière Hills volcano, Montserrat

Abstract Dome-building volcanoes, where long-term eruptive episodes can be interspersed with periods of intra-eruptive repose, are particularly challenging for volcanic hazard assessment. Defining the end of eruptive episodes is vitally important for the socioeconomic recovery of affected communities but highly problematic due to the potential for rapid transition from prolonged, seemingly low-risk repose to dangerous effusive or explosive activity. It is currently unclear what constitutes the end of repose and an eruptive episode. We show that analysis of surface deformation can characterize repose and help define an eruptive episode. At Soufrière Hills volcano, Montserrat, the long-term post–2010 deformation at 12 continuous GPS stations requires the pressure in the magma system to have increased with time; time-dependent stress relaxation or crustal creep cannot explain the deformation trends alone. Continued pressurization within the magmatic system during repose could initiate a renewed eruption, qualifying as sustained unrest and therefore continuation of the eruptive episode. For Soufrière Hills volcano, persistent magma pressurization highlights the need for sustained vigilance in the monitoring and management of the volcano and its surroundings, despite the last eruptive activity ending in 2010. Our results show promise for application to other dome-building volcanoes.

Probabilistic volcanic hazard assessment at an active but under-monitored volcano: Ceboruco, Mexico

A probabilistic volcanic hazard assessment (PVHA) for Ceboruco volcano (Mexico) is reported using PyBetVH, an e-tool based on the Bayesian Event Tree (BET) methodology. Like many volcanoes, Ceboruco is under-monitored. Despite several eruptions in the late Holocene and efforts by several university and government groups to create and sustain a monitoring network, this active volcano is monitored intermittently rather than continuously by dedicated groups. With no consistent monitoring data available, we look at the geology and the eruptive history to inform prior models used in the PVHA. We estimate the probability of a magmatic eruption within the next time window (1 year) of ~ 0.002. We show how the BET creates higher probabilities in the absence of monitoring data, which if available would better inform the prior distribution. That is, there is a cost in terms of higher probabilities and higher uncertainties for having not yet developed a sustained volcano monitoring network. Next, three scenarios are developed for magmatic eruptions: i) small magnitude (effusive/explosive), ii) medium magnitude (Vulcanian/sub-Plinian) and iii) large magnitude (Plinian). These scenarios are inferred from the Holocene history of the volcano, with their related hazardous phenomena: ballistics, tephra fallout, pyroclastic density currents, lahars and lava flows. We present absolute probability maps (unconditional in terms of eruption size and vent location) for a magmatic eruption at Ceboruco volcano. With PyBetVH we estimate and visualize the uncertainties associated with each probability map. Our intent is that probability maps and uncertainties will be useful to local authorities who need to understand the hazard when considering the development of long-term urban and land-use planning and short-term crisis management strategies, and to the scientific community in their efforts to sustain monitoring of this active volcano.

Automated Eruption Forecasting at Frequently Active Volcanoes Using Bayesian Networks Learned From Monitoring Data and Expert Elicitation: Application to Mt Ruapehu, Aotearoa, New Zealand

Volcano observatory best practice recommends using probabilistic methods to forecast eruptions to account for the complex natural processes leading up to an eruption and communicating the inherent uncertainties in appropriate ways. Bayesian networks (BNs) are an artificial intelligence technology to model complex systems with uncertainties. BNs consist of a graphical presentation of the system that is being modelled and robust statistics to describe the joint probability distribution of all variables. They have been applied successfully in many domains including risk assessment to support decision-making and modelling multiple data streams for eruption forecasting and volcanic hazard and risk assessment. However, they are not routinely or widely employed in volcano observatories yet. BNs provide a flexible framework to incorporate conceptual understanding of a volcano, learn from data when available and incorporate expert elicitation in the absence of data. Here we describe a method to build a BN model to support decision-making. The method is built on the process flow of risk management by the International Organization for Standardization. We have applied the method to develop a BN model to forecast the probability of eruption for Mt Ruapehu, Aotearoa New Zealand in collaboration with the New Zealand volcano monitoring group (VMG). Since 2014, the VMG has regularly estimated the probability of volcanic eruptions at Mt Ruapehu that impact beyond the crater rim. The BN model structure was built with expert elicitation based on the conceptual understanding of Mt Ruapehu and with a focus on making use of the long eruption catalogue and the long-term monitoring data. The model parameterisation was partly done by data learning, complemented by expert elicitation. The retrospective BN model forecasts agree well with the VMG elicitations. The BN model is now implemented as a software tool to automatically calculate daily forecast updates.

Development and persistence of hazardous atmospheres in a glaciovolcanic cave system—Mount Rainier, Washington, USA

Glaciovolcanic cave systems, including fumarolic ice caves, can present variable atmospheric hazards. The twin summit craters of Mount Rainier, Washington, USA, host the largest fumarolic ice cave system in the world. The proximity of fumarole emissions in these caves to thousands of mountaineers each year can be hazardous. Herein we present the first assessment and mapping of the atmospheric hazards in the Mount Rainier caves along with a discussion on the microclimates involved in hazard formation and persistence. Our results are compared to applicable life-safety standards for gas exposure in ambient air. We also describe unique usage of Self-Contained Breathing Apparatus (SCBA) at high altitude. In both craters, subglacial CO2 traps persist in multiple locations due to fumarole output, limited ventilation, and cave morphology. CO2 concentrations, calculated from O2 depletion, reached maximum values of 10.3 % and 24.8 % in the East and West Crater Caves, respectively. The subglacial CO2 lake in West Crater Cave was persistent, with atmospheric pressure as the main factor influencing CO2 concentrations. O2 displacement exacerbated by low O2 partial pressure at the high summit altitude revealed additional cave passages that can be of immediate danger to life and health (IDLH), with O2 partial pressures as low as 68.3 mmHg. Planning for volcanic research or rescue in or around similar cave systems can be assisted by considering the implications of atmospheric hazards. These findings highlight the formation mechanisms of hazardous atmospheres, exploration challenges, the need for mountaineering and public awareness, and the broader implications to volcanic hazard assessment and research in these environments.

Large Uncertainties Are Pervasive in Long-Term Frequency-Magnitude Relationships for Volcanoes in Southeast Asia

Frequency-magnitude relationships are a fundamental aspect of volcanic hazard and risk analysis. Typically, frequencies of previously recorded eruptions are used to identify such relationships. This works well for volcanoes that are well-studied, but it can take a long, sustained, and resource intensive research effort to compile complete eruption records. Further, the level of completeness that can be achieved will vary around the world as a function of climatic conditions, eruption style, and duration of written records. Given the importance of understanding hazard and risk for disaster risk reduction, how can frequency-magnitude relationships be determined for volcanoes with little or no eruption records? Analogue models have been used to supplement the eruption records of volcanoes with limited or no recorded eruptions. However, there has been little effort undertaken to compare the agreeability of different approaches to estimating frequency-magnitude relationships using these analogue models. This has implications for volcanic hazard and risk assessment, if different approaches are considered credible, yet yield vastly different estimates. In this work we compare frequency-magnitude relationships for volcanoes in Southeast Asia, a region where eruption records are known to be very incomplete. We do this by first reviewing published frequency-magnitude relationships to evaluate the agreeability between different approaches, and then develop a top-down multi-model Bayesian updating approach to deriving frequency-magnitude relationships for a wide variety of volcanoes in Southeast Asia (n = 176). Our review of published estimates found that there is considerable variability between published eruption probabilities for volcanoes in Southeast Asia. We also found that using different analogue models in the Bayesian analysis can lead to considerably different frequency-magnitude relationships (over an order of magnitude in some cases), highlighting the importance of using multiple models to ensure robust probability estimations are obtained. Sensitivity analysis demonstrated that the choice of model averaging or model combination method can influence the resulting frequency-magnitude estimations, whilst the choice to incorporate or remove uncertain eruption records had modest or no effect. Our findings provide important considerations for estimating frequency-magnitude relationships in volcanic hazard and risk assessments, and a method to combine multiple frequency-magnitude models whilst accounting for uncertainty in our estimations.

Probabilistic Volcanic Hazard Assessment of the 22.5–28°S Segment of the Central Volcanic Zone of the Andes

Evaluation of volcanic hazards typically focusses on single eruptive centres or spatially restricted areas, such as volcanic fields. Expanding hazard assessments across wide regions (e.g., large sections of a continental margin) has rarely been attempted, due to the complexity of integrating temporal and spatial variability in tectonic and magmatic processes. In this study, we investigate new approaches to quantify the hazards of such long-term active and complex settings, using the example of the 22.5–28°S segment of the Central Volcanic Zone of the Andes. This research is based on the estimation of: 1) spatial probability of future volcanic activity (based on kernel density estimation using a new volcanic geospatial database), 2) temporal probability of future volcanic events, and 3) areas susceptible to volcanic flow and fall processes (based on computer modeling). Integrating these results, we produce a set of volcanic hazard maps. We then calculate the relative probabilities of population centres in the area being affected by any volcanic phenomenon. Touristic towns such as La Poma (Argentina), Toconao (Chile), Antofagasta de la Sierra (Argentina), Socaire (Chile), and Talabre (Chile) are exposed to the highest relative volcanic hazard. In addition, through this work we delineate five regions of high spatial probability (i.e., volcanic clusters), three of which correlate well with geophysical evidence of mid-crustal partial melt bodies. Many of the eruptive centres within these volcanic clusters have poorly known eruption histories and are recommended to be targeted for future work. We hope this contribution will be a useful approach to encourage probabilistic volcanic hazard assessments for other arc segments.

Structure of the magma plumbing system beneath Semisopochnoi Island (Aleutian Arc) inferred from seismic tomography

Semisopochnoi Island is a remote and little-studied volcanic island in the western part of the Aleutian Arc. The existence of several active volcanic centers and a 5000–7000-year-old large caldera makes this island an important site for volcanic hazard assessment in the Northern Pacific. Based on local seismicity data recorded by six permanent seismic stations, we created the seismic tomography model, including the 3D distributions of Vp, Vs, and Vp/Vs ratios to a depth of 10 km. This model provides the first geophysical insight into the interior structure of Semisopochnoi Island and sheds light on the processes in the magma plumbing system beneath all volcanic centers on the island. At depths of 5–10 km, we observed a columnar-shaped high Vp/Vs-ratio anomaly below the caldera in the central part of the island, which likely represents the steady magma conduit. This conduit is headed by a prominent high Vp/Vs-ratio anomaly located 3–5 km directly below the caldera, which represents the magma reservoir feeding Cerberus and other Holocene-aged volcanic centers on Semisopochnoi Island.

Influence of the topography of stratovolcanoes on the propagation and channelization of dense pyroclastic density currents analyzed through numerical simulations

We applied systematically the branching energy cone model to a large ( $$N=50$$ ) set of stratovolcanoes around the world in order to evaluate the main topographic characteristics that may control the propagation of dense pyroclastic density currents (PDCs). Results indicate that the channelization efficiency of a PDC is strongly controlled by the relative scale between flow size and volcano topographic features. Most of the studied stratovolcanoes topographies are able to induce significant PDC channelization in proximal domains, while strong channelization in distal zones is mainly observed for volcanoes with steep flanks, with long, uninterrupted valleys, and with catchments zones of pyroclastic material (i.e. the valleys heads) located near the source. From the statistical analysis of numerical results, we recognise five groups of stratovolcanoes in terms of the mode of interaction between their topographies and dense PDCs: (1) intense channelization through different valleys up to distal domains (e.g. Colima and Peteroa); (2) intense channelization through a single, dominant valley up to distal domains (e.g. Reventador and Mt. St. Helens); (3) intense channelization near the source and moderate distal channelization, frequently involving intertwined drainage networks (e.g. Tungurahua and El Misti); (4) potentially intense channelization only near the source, typically involving flat distal topographies (e.g. Sinabung and Mayon); and (5) weak channelization in proximal domains, resulting in efficient early energy dissipation and thus reduced PDC run-out distance (e.g. Kelut and Akagi). The relevance of this classification lies in the possibility of defining volcanic analogues (defined here as volcanoes that share a suite of topographic characteristics and may be considered comparable to a certain extent) and identifying the main processes that may affect PDC propagation in specific topographic contexts. These aspects are useful for studying poorly documented volcanic edifices and for volcanic hazard assessment. Additionally, we compare this classification with published morphometric characteristics of volcanoes, showing morphometric parameters such as the mean slope of the low flank, irregularity index, ratio of volcano height and basal width and ratio of crater width and basal width are useful variables for recognising the groups we defined. These parameters can be used as rough indicators of the expected interaction patterns between the topography of a given volcano and dense PDCs.

VIGIL: A Python tool for automatized probabilistic VolcanIc Gas dIspersion modeLling

Probabilistic volcanic hazard assessment is a standard methodology based on running a deterministic hazard quantification tool multiple times to explore the full range of uncertainty in the input parameters and boundary conditions, in order to probabilistically quantify the variability of outputs accounting for such uncertainties. Nowadays, different volcanic hazards are quantified by means of this approach. Among these, volcanic gas emission is particularly relevant given the threat posed to human health if concentrations and exposure times exceed certain thresholds. There are different types of gas emissions but two main scenarios can be recognized: hot buoyant gas emissions from fumaroles and the ground and dense gas emissions feeding density currents that can occur, e.g., in limnic eruptions. Simulation tools are available to model the evolution of critical gas concentrations over an area of interest. Moreover, in order to perform probabilistic hazard assessments of volcanic gases, simulations should account for the natural variability associated to aspects such as seasonal and daily wind conditions, localized or diffuse source locations, and gas fluxes. Here we present VIGIL (automatized probabilistic VolcanIc Gas dIspersion modeLling), a new Python tool designed for managing the entire simulation workflow involved in single and probabilistic applications of gas dispersion modelling. VIGIL is able to manage the whole process from meteorological data processing, needed to run gas dispersion in both the dilute and dense gas flow scenarios, to the post processing of models’ outputs. Two application examples are presented to show some of the modelling capabilities offered by VIGIL.

A scenario-based volcanic hazard assessment for the Mount Meager Volcanic Complex, British Columbia

The Mount Meager Volcanic Complex (Mount Meager) is a glacier-clad stratovolcanic system in southwestern British Columbia which last erupted over 2400 years ago (VEI 4). While this is Canada’s most recent major explosive eruption, most past research on Mount Meager has focused on its numerous and large volume landslides and thus the volcanic hazard characteristics remain understudied. Here we present a suite of scenario-based hazard maps and an assessment addressing a range of potential future explosive eruptions and associated hazards. In order to overcome limited knowledge of the eruptive history, numerical models have been used to simulate the primary syneruptive hazards of concern (dome-collapse pyroclastic density currents, lahars and tephra fallout) largely utilizing eruption parameters from analogous volcanoes, i.e., glacier-clad stratovolcanoes in a subduction zone setting. This study provides a framework for similar volcanic hazard studies where geologic data is limited, funds are minimal, and access is difficult. Furthermore, this sets the stage for recognizing volcanic hazards in the Canadian landscape, providing a resource to prepare for and mitigate potential impacts well in advance of a crisis situation.

Probabilistic Volcanic Hazard Assessment for National Park Infrastructure Proximal to Taranaki Volcano (New Zealand)

Hazard assessment for infrastructure proximal to a volcanic vent raises issues that are often not present, or not as severe in hazard assessments for more distal infrastructure. Proximal regions are subject to a greater number of hazardous phenomena, and variability in impact intensity increases with the hazard magnitude. To probabilistically quantify volcanic hazard to infrastructure, multiple volcanic hazards and their effects on exposed elements need to be considered. Compared to single-hazard assessments, multi-hazard assessments increase the size and complexity of determining hazard occurrence and magnitude, typically introducing additional uncertainties in the quantification of risk. A location-centred approach, focusing on key locations rather than key hazards, can simplify the problem to one requiring identification of hazards with the potential to affect the location, followed by assessment of the probability of these hazards and their triggering eruptions. The location-centred approach is more compatible to multi-source hazards and allows for different hazard estimation methodologies to be applied as appropriate for the infrastructure type. We present a probabilistic quantification of volcanic hazard using this location centred approach for infrastructure within Te Papakura o Taranaki National Park, New Zealand. The impact to proposed park infrastructure from volcanic activity (originating from Mt. Taranaki) is quantified using a probability chain to provide a structured approach to integrate differing hazard estimation methods with eruption probability estimates within asset lifetimes.This location-centered approach provides quantitative estimates for volcanic hazards that significantly improve volcanic hazard estimates for infrastructure proximal to the Taranaki summit vent. Volcanic mass flows, predominantly pyroclastic surges or block and ash flows, are most likely (probability >0.8) to affect walking tracks if an eruption occurs. The probability of one or more eruption(s) in the next 50 years is estimated at 0.35–0.38. This use of probability chains and a location centered assessment demonstrates a technique that can be applied to proximal hazard assessments globally.

Reconstruction of the subaerial Holocene volcanic activity through paleomagnetic and 14C dating methods: El Hierro (Canary Islands)

Volcanic hazard assessment relies on the accurate knowledge of the eruptive style and recurrence of volcanic eruptions in the past. At El Hierro (Canary Islands) historical and prehistorical records are still poorly defined, and although the island was the location of one of the most recent eruptions (La Restinga, 2011 CE) of the Canarian archipelago, the recent subaerial volcanism is still poorly studied. Information about the age of Holocene volcanic activity as well as the stratigraphy of the deposits is scarce: few eruptions are dated so far, whereas the others are classified as pre-or Holocene events considering lava flow characteristics along the coast.Here, we report on the dating of eleven (Mña Chamuscada, Mña del Tesoro, Orchilla, Las Calcosas, Mña Negra, Lomo Negro, Below Lomo Negro, Cuchillo del Roque, Malpaso Member, and Mña del Guanche) Holocene subaerial eruptions, distributed along the three rift zones, combining paleomagnetic and 14C methods. We also provide geochemical analyses for nine of them. Results indicate that Mña Chamuscada and Mña del Tesoro occurred more recently than previously considered, setting them within the last two thousand years. Conversely, paleomagnetic and 14C ages found for Lomo Negro eruption are consistent with literature data (Villasante-Marcos and Pavón-Carrasco, 2014) and constrain the occurrence of this event in the XVI century CE. Finally, for Malpaso Member deposits, the two 14C datings obtained by charcoals found below and above the trachytic layer set the eruption during the Holocene epoch, between ~7300 BCE and ~4700 BCE. For the other eruptions, in two cases (Orchilla and Las Calcosas) many possible time windows during the last 14 ka have been found, whereas a few possible ages have been obtained for the others. On the whole, the resulting chronological reconstruction of the recent activity of El Hierro indicates that eruptions occurred unevenly along the three main rifts, with nine eruptions in the WNW rift, six in the NE rift, and four in the SSE rift. We document at least two periods characterized by high eruptive frequency: an old one, between 8000 BCE and 1000 BCE, with eight eruptions, three of which characterized by more evolved compositions (phonotephrite and trachyte), and a recent one, between 1000 BCE and present day, with at least seven eruptions, mainly showing basanite compositions. The new data yield a significant improvement of Holocene eruption chronology, thus are instrumental for a correct evaluation of the volcanic hazard at El Hierro.

The Volcanic Hazards of Jan Mayen Island (North-Atlantic)

Hazard assessment of remote volcanic islands provides many challenges compared to other volcanoes and volcanic fields. Here we present the first systematic volcanic hazard assessment of Jan Mayen Island, a remote island located in the North-Atlantic Ocean and home to the northernmost active subaerial volcano in the world (Beerenberg Volcano), and we discuss some of the challenges and characteristics of performing a volcanic hazard assessment of a remote volcanic island. Jan Mayen has had at least five eruptions since its discovery at the start of the 17th century. Its Holocene volcanism is mainly characterized by eruptions with styles ranging from Hawaiian to Strombolian, but also by lava domes and Surtseyan eruptions. Based on field data, remote images, topographic data, past data, and computer simulations, our study evaluates the spatial probability of new vents opening, estimates eruption recurrence rates, simulates various eruption scenarios, and produces hazard maps for the different scenarios. This work shows where the hazards of ash fall, and lava flows are more likely to affect the built infrastructure on Jan Mayen Island. This hazard assessment will assist emergency planning and the determination of future land use on the island.

The 2011 Cordón Caulle eruption triggered by slip on the Liquiñe-Ofqui fault system

Determining the mechanisms that promote large silicic eruptions is one of the biggest challenges in volcanic hazard assessment. The 2011-2012 Cordón-Caulle eruption in Chile was one of the largest silicic eruptions of the 21st century and was characterized by a rapid change from explosive to effusive behavior. This eruption was preceded by inflation from 2007 to 2009, followed by two years of barely any ground deformation. Despite intensive monitoring by geodetic and seismological data, its trigger remains undetermined. Here, we benefit from SAR imagery over the Puyehue Cordón-Caulle Volcanic Complex acquired by ALOS-1, ENVISAT and SENTINEL-1 data, to analyze the temporal and spatial behavior of ground displacements before, during and after the eruption. First, we find that a similar prolate spheroidal source explains the ground deformation for the pre-eruptive and post-eruptive periods. Then, we use 3D numerical elasto-plastic models to assess the failure conditions resulting from the pre-eruptive magma injection. Our results show that such a magma injection was too small to trigger the eruption. Therefore we explore other eruption triggers. Analytical elastic inversion models show that the ground displacements observed during the explosive phase may have been produced by slip motion along a NNW-striking dextral-strike slip, double-branch fault of the north-trending Liquiñe-Ofqui Fault System (LOFS), or along a single southern branch fault of the LOFS and collapse of the caldera. When investigating the elasto-plastic deformation pattern resulting from dextral slip along this branch-fault system, we obtain a sub-vertical dilatational plastic zone that connects the reservoir wall to the surface in a location that coincides with that of the 2011 eruption. Hence, we propose that this LOFS branch-fault eventually destabilized (perhaps weakened by the 2007-2009 episode of magma injection), and then slipped in a way that opened channels for fluid migration from the magma reservoir up to the surface.

Toward Autonomous Detection of Anomalous GNSS Data Via Applied Unsupervised Artificial Intelligence

Artificial intelligence applications within the geosciences are becoming increasingly common, yet there are still many challenges involved in adapting established techniques to geoscience data sets. Applications in the realm of volcanic hazards assessment show great promise for addressing such challenges. Here, we describe a Jupyter Notebook we developed that ingests real-time Global Navigation Satellite System (GNSS) data streams from the EarthCube CHORDS (Cloud-Hosted Real-time Data Services for the geosciences) portal TZVOLCANO, applies unsupervised learning algorithms to perform automated data quality control (“noise reduction”), and explores autonomous detection of unusual volcanic activity using a neural network. The TZVOLCANO CHORDS portal streams real-time GNSS positioning data in 1[Formula: see text]s intervals from the TZVOLCANO network, which monitors the active volcano Ol Doinyo Lengai in Tanzania, through UNAVCO’s real-time GNSS data services. UNAVCO’s real-time data services provide near-real-time positions processed by the Trimble Pivot system. The positioning data (latitude, longitude and height) are imported into the Jupyter Notebook presented in this paper in user-defined time spans. The positioning data are then collected in sets by the Jupyter Notebook and processed to extract a useful calculated variable in preparation for the machine learning algorithms, of which we choose the vector magnitude for further processing. Unsupervised K-means and Gaussian Mixture machine learning algorithms are then utilized to locate and remove data points (“filter”) that are likely caused by noise and unrelated to volcanic signals. We find that both the K-means and Gaussian Mixture machine learning algorithms perform well at identifying regions of high noise within tested GNSS data sets. The filtered data are then used to train an artificial intelligence neural network that predicts volcanic deformation. Our Jupyter Notebook has promise to be used for detecting potentially hazardous volcanic activity in the form of rapid vertical or horizontal displacement of the Earth’s surface.

Calibration strategies of PDC kinetic energy models and their application to the construction of hazard maps

The availability of computer tools able to describe the behavior of pyroclastic density currents (PDCs) with uncertainty quantification is of primary importance for the assessment of volcanic hazard. A common strategy to assess the intrinsic variability of these phenomena is based on the analysis of large sets of numerical simulations with variable input parameters. The use of models fast enough to allow for a large number of simulations, such as the so-called kinetic energy models, is thus advantageous. Due to the sensitivity of kinetic energy models to poorly constrained input parameters, the definition of their variation ranges is a critical step in the construction of hazard maps and a numerical calibration becomes necessary. We present a set of reproducible and structured calibration procedures of numerical models based either on a reference deposit or on the distribution of runout distance or inundation area of documented PDCs. In the first case, various metrics can be adopted to compare the model results with the reference PDC deposit (root mean square distance, Hausdorff distance, and Jaccard index), facilitating the development of scenario-based hazard assessments. Calibrations based on the distribution of runout distance or inundation area allow the construction of probabilistic hazard maps that are not conditioned on the occurrence of a specific scenario, but rather reflect the variability of the documented PDCs during the time window considered. Importantly, our calibration strategies allow one to set the input parameters considering their potential statistical dependence. These procedures have been implemented on the user-friendly versions of two kinetic energy models: ECMapProb 2.0 and BoxMapProb 2.0, whose functionalities are presented for the first time in this paper. The different calibration strategies and the functionalities of the two programs are illustrated by considering three case studies: El Misti (Peru), Merapi (Indonesia), and Campi Flegrei (Italy).