Final Report Summary - VUELCO (Volcanic unrest in Europe and Latin America: Phenomenology, eruption precursors, hazard forecast, and risk mitigation)
Executive Summary:
Volcanic unrest is a complex multi-hazard phenomenon of volcanism. Uncertainties surrounding cause and outcome of unrest have substantial implications for short-term volcanic hazard assessment. The knowledge of the causative links between subsurface processes, resulting unrest signals and imminent eruption is, today, inadequate to deal effectively with crises of volcanic unrest. This results predominantly from the uncertainties in identifying the causative processes of unrest and as a consequence in forecasting its short-term evolution. However, key for effective unrest risk mitigation and management is the early and reliable identification of changes in the subsurface dynamics of a volcano and their assessment as precursors to an impending eruption.
The VUELCO research consortium has worked over the past 4 years to significantly improve our understanding of the processes behind volcanic unrest and the ability to forecast its outcome aiding decision-making and management in an unrest situation. VUELCO has tackled outstanding problems surrounding volcanic unrest via an international multi-disciplinary consortium, combining fundamental research into causes and effects of volcanic unrest with uncertainty assessment and probabilistic forecasting to improve communication, decision-making and management during volcanic unrest.
The project findings have (i) improved the mechanistic understanding of subsurface processes triggering volcanic unrest, (ii) helped identify reliable precursors associated with specific subsurface processes, (iii) improved the forecasting capacity of the outcome of volcanic unrest in the presence of scientific uncertainty, (iv) improved the capacity for early-warning and management of evolving volcanic crises beginning with monitoring capacity through hazard assessment and risk governance to decision-making and (v) improved the preparedness to cope with consequences and potentially adverse outcomes of volcanic unrest.
Project Context and Objectives:
Volcanic unrest is a complex multi-hazard phenomenon of volcanism. The fact that unrest may, but not necessarily must lead to an imminent eruption contributes significant uncertainty to short-term hazard assessment of volcanic activity world-wide. Although it is reasonable to assume that all eruptions are associated with precursory activity of some sort, the knowledge of the causative links between sub-surface processes, resulting unrest signals and imminent eruption is, today, inadequate to deal effectively with crises of volcanic unrest. This results predominantly from the uncertainties in identifying the causative processes of unrest and as a consequence in forecasting its short-term evolution. However, key for effective risk mitigation and management during unrest is the early and reliable identification of changes in the subsurface dynamics of a volcano and their assessment as precursors to an impending eruption.
When a volcano develops from dormancy through a phase of unrest important scientific, political and social questions need to be addressed. Scientific questions relate to the processes behind unrest and their associated surface signals (if any), their future evolution and their significance as precursor for eruptive phenomena. Political and social questions are related to: (i) how epistemic and aleatory uncertainties surrounding data and processes are employed in the light of probabilistic assessment of outcomes of unrest periods, (ii) how these uncertainties are communicated and (iii) how evolving crises are best managed depending on the forecasted scenario.
The overarching aim of VUELCO is to significantly improve our understanding of the processes behind volcanic unrest and the ability to forecast its outcome aiding decision-making and management in an unrest situation. The working hypothesis behind VUELCO is that current uncertainties surrounding volcanic unrest episodes and can be minimised through a concerted multi-disciplinary and cross-boundary effort.
VUELCO focuses on meeting the objectives via an interdisciplinary consortium drawing from expertise in both Europe and Latin America in three main Project Areas:
1. The research into the mechanistic processes behind unrest and their link to surface observables including capacity building for monitoring and interpretation of unrest signals (Area 1: Science of Volcanic Unrest),
2. The development of tools and protocols for uncertainty assessment and probabilistic forecasting during unrest crises (Area 2: Uncertainty Assessment and Probabilistic Forecasting),
3. The development of risk mitigation programmes, communication protocols and management tools amid scientific uncertainty (Area 3: Communication, Decision-Making and Management).
VUELCO combines fundamental research into causes and effects of volcanic unrest with uncertainty assessment and probabilistic forecasting to improve communication, decision-making and management during volcanic unrest.
The objectives of the project are:
• To improve the mechanistic understanding of subsurface processes triggering volcanic unrest.
• To identify reliable precursors associated with specific subsurface processes.
• To improve the forecasting capacity of the outcome of volcanic unrest periods in the presence of scientific uncertainty.
• To improve the management of evolving volcanic crises beginning with monitoring capacity through hazard assessment and threat analysis to decision-making.
• To improve the preparedness to cope with consequences and potentially adverse outcomes of volcanic unrest.
• To improve the capacity for early-warning of phenomena associated with and the consequences of unrest.
Project Results:
WP3
1. Most volcanic eruptions are preceded by input of less evolved, volatile-rich magma into a shallow-seated magmatic system resulting in magma convection and mixing. Detailed geochemical and petrological investigation at Campi Flegrei, Italy, illustrate several aspects of this relevant process, showing that, although, virtually all eruptions are preceded by magma mixing, not all mixing events lead to an eruption.
2. Numerical simulations of magma convection and mixing in geometrically complex volcanic systems reveal the dynamics of the processes and shed light on the geophysical signals diagnostic of such critical occurrences. Ultra-long-period seismicity is identified as the most indicative signature of on-going magma convection and mixing at depth. Transient events in the frequency band 10-4 – 10-2 Hz, suggested by the numerical simulations, are found from the analysis of monitoring records from strain meters at Campi Flegrei in heightened unrest periods; these are interpreted as the first evidence ever found of magma recharge events at depth during volcanic unrest.
3. 2D and 3D numerical models developed to characterise the influence of the regional stress field on volcanic systems have revealed the relationships between the stress configuration in the lithosphere at different scales (i.e. local, regional and plate-scale) and the distribution of volcanism, particularly in monogenetic volcanic fields. There is a potential for such methods to allow discriminating between volcanic and tectonic controls volcanic unrest. The models developed in the project are of general applicability and can be used to define the stress characteristics of many active volcanic areas, helping interpret unrest episodes.
4. Modelling of magma ascent along volcanic conduits applied to Soufrière Hills volcano on Montserrat (UK), suggests that the observed long-period seismicity can reflect brittle failure of magma at the conduit boundaries. Changes in seismicity patterns might therefore be used as early indicators of changes in ascent dynamics and eruptive styles/impacts.
5. Geodetic modelling at Soufrière Hills reveals that in spite of the current absence of eruptive activity, the eruptive cycle that started in the mid-1990s should not be deemed finished, since the system shows clear signs of over-pressurization. The origin of the over-pressurization is still not clear, and might be due to crystallization induced-degassing (second-boiling) in response to magma cooling. Consequently, the possibility of magma chamber cooling cannot be fully discounted. Extensive computer modelling and new interpretation of the on-going deformation at Montserrat has revealed that a pressurising magma body centred at 6km depth can explain the apparent uplift and lateral deformation. The new results take both topography and the stretching of the tectonic plate into account. This reduces the amount of pressurisation to such a degree, that cooling and crystallisation of the existing magma body is a viable explanation with no or decreased magma influx.
6. A high-precision, two-year long record from White Island, New Zealand was used to test if the movement of volcanic fluids like magma could be detected in a gravity record. While final results of this investigation are still pending, we can conclude that gravity changes in volcanic systems are most likely caused by shallow changes in the hydrothermal system and therefore mask potential magma movement to such a degree that conclusive interpretations of temporal gravity anomalies are very difficult.
7. A novel paradigm for volcano deformation modelling has been developed. The new framework can handle deformation sources of arbitrary geometry that are embedded in realistic rock media characterized by structural heterogeneities, temperature dependent viscoelastic response and complex topography. This approach allows for more accurate estimates of magma reservoirs locations and sizes, providing a substantial improvement in the identification of the possible sources of volcanic unrest. The new generation of forward and inverse models has been successfully applied to Cotopaxi volcano (Ecuador) to identify the magmatic source involved in the unrest phase of 2002-2003; the same modelling has provided insights into the recent eruptive phase (August 2015 onwards). Other applications in the project include unrest deformation in the Altiplano region of Bolivia, Aira caldera, Japan and Soufrière Hills volcano, Montserrat.
8. Numerical simulations of aquifer dynamics and relationships with volcanic activity suggest that the fault system can play a controlling role on heat transport within shallow hydrothermal reservoirs, therefore influencing substantially heat flow measurements at the surface. Results from Campi Flegrei highlight a dominant role of the permeability distribution, with faults representing local maxima in permeability. Simulations of magmatic unrest show that a single magmatic input event can result in an oscillatory response of the hydrothermal system recorded in geochemical (heat flux from the ground) and geophysical (deformation and gravity) signals, as a consequence of complex interactions between gas generation at depth and recharge–discharge dynamics of the reservoir. Investigating the thermal consequences of non-eruptive magma intrusions into glaciated volcanoes as exemplified by Cotopaxi volcano (Ecuador), shows that spatio-temporal variations in rock permeability are the prime contributor to rapid surface heat flux changes, with the potential to produce hazardous ice melt and flooding events without eruptive activity.
9. A comprehensive analysis of decades-long geophysical and geochemical datasets of unrest has been performed at Campi Flegrei in order to explore the capabilities of advanced multi-component gas–melt equilibrium modelling and highly accurate isotopic data to provide the ground for global interpretation of the overall processes subtending volcanic unrest. The results highlight the complexities deriving from the interplay between different degassing sources represented by magma emplaced at different depths below the volcanic system. Such a comprehensive approach demonstrates that complex patterns of geochemical and geophysical quantities recorded at unrest volcanoes can be interpreted in a unique frame, when the competing roles of the different portions of spatially extended, geometrically complex underground magmatic systems are accounted for.
10. Interrogating novel numerical fluid flow models, which simulate fluctuating hydrological recharge in volcanic environments, enabled us to quantify resultant mass redistribution and its effect on geophysical volcano monitoring signals. The results study highlight that mass storage (saturation) changes beneath a survey point, within the unsaturated zone, can generate measurable gravity changes. For a tropical climate inter-annual recharge variations can dominate over seasonal signals and spatially heterogeneous hydro-geological conditions can impact on the accuracy of relative gravity surveys. The additional loading fluctuations associated with saturation variations in the unsaturated zone may also have implications for other geophysical techniques such as monitoring of ground deformation. Ultimately the findings inform the derivation of baseline behaviour of volcanoes in order to better identify pre-eruptive unrest behaviour.
WP4
1. We experimentally determined pre-eruptive processes and magma storage conditions at Tungurahua (Ecuador), Layou & Roseau (Dominica) using a combination of experimental (laboratory-based) and analytical (glass inclusion-based) approaches.
2. We analysed tephra from Mourne aux Diables (Dominica) eruptions in terms of bulk-rock chemistry and mineralogy.
3. We developed an experimental crystal-liquid equilibria database for Cotopaxi Ecuador and applied it to define conditions of the 2015 magmatic unrest at the volcano.
4. We developed mineralogical approaches to constrain pressure-temperature conditions, mechanisms and timescales of basalt magma ascent.
5. We performed experimental simulations of conduit processes (ascent and degassing) for volatile-rich basaltic magmas and found evidence for an anomalous behavior of CO2 and the generation of CO2-supersaturated melts during ascent.
6. We successfully replicated natural glass volatile (H2O, CO2, S) degassing patterns using experiments and showed that gas-melt chemical disequilibrium may occur during magma degassing.
7. We composed a robust compositionally-dependent database for the Raman spectra of natural silicate glasses along the calcalkaline series. In addition, an empirical model based on both the acquired Raman spectra and an ideal mixing equation between basaltic and rhyolitic end-members was constructed.
8. Using magma mixing experiments we developed an efficient mechanism for hybridisation of magma, whereby bubbles originating from a deep mafic melt intruded into a shallow felsic melt.
9. We discovered that the experimentally validated physics of densification in magmas can be divided into two regimes and we derived governing equations for both of them: (a) surface-tension dominated densification, and (b) buoyancy dominated densification.
10. We were able to successfully predict the rates of densification for volcanic systems I support of observations that indicate that magma densification occurs at all volcanoes.
11. We have built a phase diagram that can be interrogated to predict whether or not densification of a system of volcanic particles or bubbles will lead to pressure increases leading to catastrophic fracturing or not.
12. Conditions and processes of phreatic and hydrothermal explosions were investigated experimentally for several targets (Campi Flegrei, Dominica, White Island), increasing our understanding of key aspects of this very variable eruption type.
13. Even small phreatic explosions involving fine-grained ash fall deposits, affected by alunitic alteration, produce a significant and much larger than expected amount of very fines (< 10 µm) with a high potential to instigate chronic diseases.
14. Depending mainly on fluid and rock composition and especially the pH, alteration may both hinder and favour the degassing of hydrothermal fluids and thus the risk of phreatic explosions.
15. Geochemical experiments were successfully tested as a tool to validate reactive transport simulations in volcanic settings. For Montserrat, we were able to fingerprint the response of groundwater to magmatic perturbation.
WP5
1. We have developed and proposed a definition of volcanic unrest: “ The deviation from the background or baseline behaviour of a volcano towards a behaviour which is a cause for concern in the short-term because it might prelude an eruption.”
2. We have performed a global audit of volcanic unrest and created a database from the reported information on unrest at 228 volcanoes. The data is categorised into pre-eruptive or non-eruptive unrest indicators at four different subaerial volcano types and submarine volcanoes as defined by the GVP. Unrest timelines demonstrate how unrest evolved over time and highlight different classes of unrest including reawakening, pulsatory, prolonged, sporadic and intra- eruptive unrest. Statistical tests indicate that pre-eruptive unrest duration was different across different volcano types. 50% of stratovolcanoes erupted after about one month of reported unrest. At large calderas this median average duration of pre-eruptive unrest was about twice as long. At almost five months, shield volcanoes had a significantly longer unrest period before the onset of eruption, compared to both large calderas and stratovolcanoes. At complex volcanoes, eruptive unrest was short lived with only a median average duration of two days. We find that there is only a poor correlation between the length of the inter-eruptive period and unrest duration in the data; statistical significance was only detected for the pair-wise comparison of non-eruptive unrest at large calderas and stratovolcanoes. Results indicate that volcanoes with long periods of quiescence between eruptions will not necessarily undergo prolonged periods of unrest before their next eruption. The resultant database is publically available.
3. We have developed and populated a Dynamic Data Base aimed to process, interpret, and identify seismic precursors that may be causally related to eruptions producing columns of over 4 km height above the crater of Popocatepetl volcano in Mexicos since 1997. The database is available for interrogation and has been incorporated into the WOVOdat master database on volcanic unrest.
4. Multi-parameter unrest data from the El Hierro volcanic crises in 2011/12 has been incorporated into the WOVOdat master database.
WP6
1. An overview was compiled outlining the monitoring techniques already present on each of the target volcanoes. This information was obtained via questionnaires, which were completed by monitoring specialists at each of the volcanoes. This information was then summarised as factsheets for each. This allowed us to identify those techniques with the potential to forecast periods of unrest. It also enabled us to identify the gaps in the existing monitoring network, which might need addressing.
The overview of monitoring techniques already available at the target volcanoes will guide future recommendations for the minimum instruments and processing tools required at each volcano for short-term unrest forecasting.
2. We identified accelerating data sequences within the seismic records and applied the Failure Forecast Method (FFM) as a forecasting tool. The significant result from this is the entire seismic signal must be broken down into the individual waveform components as defined by their cross-correlations. These systems behave according to the FFM and can be used as forecasting candidates. Using the cross-correlation technique in conjunction with FFM provides the most accurate fit of the least squared linear regression of the data.
3. We applied the FFM to an unrest period observed at Chiles and Cerro-Negro volcano, Ecuador. Analysis identified clusters of similar seismic events, but with no change in amplitude through time. This seismicity was interpreted as generated by hydrothermal activity due to the contact of ascending magma with aquifers.
4. By integrating a seismic trace in a routine process, VLPs can be easily identified and later individually processed by removing the instrument response. This allows the detection of ground motion within a frequency band too high, and too small in amplitude for GPS measurements, and too low in frequency to be detectable through usual seismic monitoring procedures. VLPs may play a major role in outgassing events and other sudden movements of volcanic fluids.
5. The numerical integration of seismic traces has now been adopted as a routine processing tool at the MVO on Montserrat and IGENP in Ecuador. Hence, VLP signals can be easily identified and later individually processed and interpreted by removing the instrument response.
6. Through forward modelling, we obtained an entire set of complex volcanic source characteristics in seismic moment tensors that can be used to identify volcanic seismic unrest for different types of seismic signals.
7. We combined tilt and seismic interpretation of Tungurahua activity: Numerical modelling of the monitored tilt signal of Tungurahua volcano, Ecuador, has shown that the signal cannot be explained by the usual inflation/deflation cycles of the volcanic edifice. Pressures necessary to satisfy the data would be unreasonably high. In addition, the fact that explosions on this volcano are often preceded by a ‘deflation’ episode, and therefore depressurization, makes an interpretation difficult. We have demonstrated that an alternative explanation of shear stress induced tilt can explain both the timing of the tilt signal and its amplitude. This will have huge implications for using the tilt signal together with the seismicity as a forecasting tool on Tungurahua and other silicic volcanoes worldwide.
8. SO2, CO2, H2O, HCl and H2S are the most frequently measured gases in volcanic settings and provide a direct insight in magma composition and its temporal evolution. Direct measurements at fumaroles are increasingly replaced by in-plume measurements through ground-based spectroscopic methods in the ultraviolet to infrared spectral range (DOAS, COSPEC, FTIR). We made progress by addressing the weakest point of these measurements concerning the processing step between SO2 snapshot in the plume and calculation of SO2 flux to which a precise determination of wind speed is necessary. We developed an approach using numerical weather models in combination with meteorological data to calibrate local weather stations to estimate the plume transport speed.
WP7
1. We developed a novel prototypal probabilistic forecasting scheme (BET_UNREST) for a Bayesian Event Tree to forecast the probability of occurrence of different hazardous phenomena related to volcanic unrest, both of magmatic and non-magmatic origin. The Bayesian nature of the event-tree permits, on the one hand, to fully account for the uncertainties involved (both of aleatory and epistemic), and on the other hand, to consider all the heterogeneous sources of information: from conceptual and volcanological models to monitoring protocols, the incomplete geological record, and expert opinions. The output of the proposed Bayesian event tree (i.e. the forecast of the probability of different unrest outcomes of magmatic and non-magmatic nature) may support a full management of volcanic crisis, accounting for all the different aspects of volcanic unrest, in particular considering the risk posed by non-magmatic events, such as phreatic eruptions. Such threats are almost always underestimated, but they may pose a major risk during volcanic unrest, as documented by the 2014 Ontake (Japan) phreatic explosion, which claimed several dozen lives. This development represents a key novelty in the field of volcanic hazard assessment, which can aid risk mitigation efforts.
2. The application of Bayesian Event Tree to track the evolution of volcanic unrest at 5 of the 6 target volcanoes of the project, both real-time during the simulation experiments carried out during VUELCO (Colima, Campi Flegrei, Cotopaxi and Dominica), and “offline” for a specific target period (Popocatepetl, April to June 2013). As regards the management of the hazardous phenomena related to non-magmatic unrest, the application of the new Bayesian Event Tree to the Dominica volcanic island, characterized mostly by hydrothermal activity, has allowed the practical usefulness to account for non-magmatic hazardous outcomes in a volcanic unrest crisis.
3. Among the aforementioned applications of the Bayesian Event Tree for managing the volcanic unrest, we highlight the one at Campi Flegrei, as it is one of the highest volcanic risk areas in the world. The event tree provides a real-time estimation of the daily to weekly probability of eruption of different sizes and from different possible vent positions, as a function of the real-time monitoring at the volcano. In a fruitful synergy with another FP7 project (MED-SUV), the results from VUELCO application at Campi Flegrei have been coupled to MED-SUV simulations of tephra fallout, to estimate the short-term tephra fallout hazard, accounting for the wind forecast of the next days. This represents a cutting-edge tool to bridge the scientific knowledge of volcanic unrest and the decision-making purposes.
4. The several experiments carried out during the VUELCO projects highlight clearly the state-of-the-art and strengths and weaknesses of the current interface between volcanologists and emergency managers. This has allowed us to research this link interface in greater detail and we have proposed a general framework based on the hazard/risk separation principle that allows volcanologists and emergency managers to clarify their role and responsibilities in the whole decision-making process.
WP8
1. Using a specifically developed questionnaire, we have performed a global audit of existing communication protocols and practices during volcanic crises. Responses received from stakeholders informed the preparation of best–practice guidelines.
2. The resultant best-practice recommendation and policy-guidelines informed by both the global audit, research conducted in the project and direct interaction with emergency managers and volcano observatory personnel during unrest simulation exercises, aim at developing more efficient and effective inter and intra-stakeholder communication for improved decision-making and management of unrest events.
WP9
1. We audited the level of trust in local authorities and opinion makers around Cotopaxi volcano using questionnaires and field interviews. This set of question became a part of the Insecurities Methodology that has been promulgated to VUELCO partners. General findings include:
• The trust of the Ecuadorian population in authorities is directly affected by broader security concerns. Those authorities which are seen as having a responsibility for every-day security issues, such as crime prevention, are accorded low levels of over-all trust, and this negatively effects the population’s trust in the authorities with respect to a potential volcanic hazard.
• Despite relatively high levels of religiosity, measured with the proxy of church attendance, church leaders are not seen as being particularly trust-worthy in relation to advice on disaster preparedness and response.
• In spaces where scientific expertise is accorded higher levels of trust than government officials, care needs to be taken to ensure that scientists (and their organizations) do not become associated with negative opinions of the broader government.
2. We performed an audit into volcanic risk related decision-making processes in different countries to understand how systems are organized and work. work and are organized. The audit was based on information provided by decision-makers from Mexico, Dominica (W.I.) La Reunion (France), Italy, Ecuador, Canary Island (Spain), Montserrat (UK), and El Salvador.
Key findings included:
• Generally there is a low volcanic risk perception of the general public, even in the large towns close to volcanoes. Decision-makers deem it important that people participate in drills, information campaign, etc.
• Although hazard maps and emergency plans have been generally drawn up (although not regularly updated), exercises and information campaigns have been carried out only in a few countries and for a small amount of people (even in the most exposed populated areas).
• A cost/benefit analysis regarding mitigation efforts is generally not applied.
• Different entities (from Mayor to Ministry) are involved in all levels of communication. Language used is often the official national one only, although sometimes local dialects are also used.
• In order of priority, decision-makers want to know from scientists: 1. kind of expected hazardous phenomena, 2. probability of occurrence of different possible hazardous phenomena, 3. areas with different hazard levels, 4. evolution and duration of the phenomena, 5. time needed to adopt each mitigation action.
• Low importance seems to be given instead to the possibility of false alarms and associated collateral risks. The evaluation of the costs of possible mitigation actions is generally neglected .
• Timeliness of the information appears to be preferred over their accuracy.
3. Prior to the commencement of field work in Ecuador, and before the simulation exercise in Colima, Mexico, a detailed survey of the political and social history of the states was conducted, in particular with relation to security issues. The histories were integrated into the development of the Insecurities Methodology, and informed several of the survey questions. The past history of social unrest in Mexico was also used to inform the audit of the Colima Simulation Exercise. The social history of unrest is a crucial variable in understanding community perceptions of natural hazards and their risk awareness.
4. Social unrest has a direct impact on the perceptions of the state’s disaster preparedness and response mechanisms. Institutions, such as the police, that are seen as a source of insecurity will be accorded lower levels of trust in relation to disaster response. Communication strategies to prepare communities for potential disasters need to take into account the lived experiences of communities. Arguing that a natural hazard should be a central concern in the face of personal and community insecurity will not resonate and may erode the confidence of communities in the disaster management regime.
5. The application of theories of securitization and human security has provided an useful addition to the social vulnerability literature on disaster risk reduction (DRR). It has put into focus the ways in which DRR strategies can increase, rather than decrease, perceptions of insecurity. As the post-simulation audit of the Colima exercise indicates, the state strategy of moving displaced peoples into secure compounds where the freedom of movement is restricted, and family life is disrupted through communal living, if put into practice, is likely to reduce compliance of community members.
6. Significant recent court cases on natural risk governance, including L’Aquila, are framed by process of deduction within a generalised legal infrastructure in order to identify the root causes of the apparent status quo of risk governance. We developed a forensic approach not only to identify the legal responsibilities of societal risk managers and the managerial risks that they face and their causes, but also to consider possible mitigation strategies.
We identify the critical issue of managerial risk vulnerability related to ‘standard equivocality’, which is the absence of commonly recognised standards for hazard communications to risk decision makers. This absence may result from the lack of regulation of relevant practices and practitioners. We offer some recommendations to fuel debate not only within those science groups that reacted to the L’Aquila case but also the scientific community as a whole. Finally, we argue that checklists developed in the framework of VUELCO represent a rational and methodical way to develop acceptable practice standards focussed upon the difficult risk mitigation choices that are made by civil protection authorities and at-risk individuals.
7. We have reviewed the literature on how scientific output from the volcanological community is used in decision-making and we have developed guidelines that may allow scientists and decision makers to bridge the gap between volcanological knowledge and a rational decision making, based on the Cost-Benefit analysis (CBA) of risk mitigation. Basically, CBA is a quantitative study of the trade-offs involved in making a selected mitigating action in the interests of public safety, or of any other stakeholder. As the probability of a specific event exceeds a threshold, defined through CBA, some commensurate mitigation action may become worth taking. This procedure constitutes a crucial link between science and society, establishing rational and transparent decision-making procedures, and clarifying the responsibility of each partner involved in managing a crisis. We have applied the CBA in an illustrative (retrospective) example showing how it could have been used during the VUELCO simulation exercise at Campi Flegrei to minimise the risk posed by pyroclastic density currents. This has been achieved by including results of pyroclastic density currents propagation performed in the FP7 project MED-SUV. This retrospective application to Campi Flegrei can be seen as a first simple and preliminary application.
8. The four volcanic unrest simulation exercises conducted during the project (Colima, Campi Flegrei, Cotopaxi, Dominica) highlighted a number of strengths and weakness including those within: (1) monitoring systems; (2) geological and structural models; (3) local Civil Protection scientific advice systems and emergency management arrangements; (4) communication procedures among stakeholders and; (5) alert level systems.
Based on audits of the simulation exercises we were able was to provide local institutions (scientific, operational and governmental) with important clues for possible improvements in the above-mentioned fields.
The exercises facilitated an extensive assessment of the delicate interactions between scientists and decision makers and the identification of critical relationships and vulnerabilities. They also made possible the testing of probabilistic forecasting models and elicitation methods in simulated “real time” within risk assessment procedures.
The participation of representatives of civil protection authorities from many countries, including those not involved directly in the VUELCO project, was also very important. It encouraged the sharing of expertise, experience and good practices.
In Mexico there was the opportunity to involve the local population in educational activities and a complex evacuation drill, and thereby measure their preparedness and the utility of mitigation measures.
We were able to identify a number of possible enhancements that, if put in place, could really improve the civil protection response to different kinds of emergencies and not only those caused by volcanoes.
Based on audits of the performed simulations, we have devised guidelines to inform the design of future volcanic unrest simulation exercises with global application.
Potential Impact:
Impacts:
In addition to the impact created by dissemination activities of the S&T developments in the project (including conference presentation, direct non-scientific stake-holder interaction, workshops, summer schools, public relations material, press releases, press coverage, virtual agora, peer-reviewed journal publications, video, tweets, peer-to-peer messaging (eg., FACEBOOK)) a number of impacts of the project are specifically highlighted below.
Wider dissemination activities within the volcanological community and beyond included the organisation of many scientific sessions at major international conferences such as EGU, AGU, IAVCEI and IUGG. Scientific papers have been and will continue to be published after the end of funding in high-impact journals.
1. A suite of next-generation physico-mathematical (numerical) models capable of reproducing several aspects of the complex dynamics of volcanic unrest and pre-eruptive processes has been developed and/or refined and improved. Together, these models provide the state-of-the-art reference for the quantitative interpretation of geophysical and geochemical signals recorded at restless volcanoes in terms of magmatic movements, interaction between different magmatic sources, relationships between volcanic and regional/tectonic stress fields, magma degassing, rock properties, presence and dynamics of aquifers. A cook-book to introduce novice modellers to and to validate code from geodetic finite element modelling is available by ‘open access’.
2. Applications of the newly developed models at the target volcanoes and beyond volcanoes highlight the potential of the new modelling resources for new interpretation of the sources behind past and current unrest periods with a potential for a paradigm shift in how to interpret unrest signals.
3. All of the modelling capabilities developed are general and can be applied to any restless volcano, thus, expanding the impacts expected by the mathematical modelling activities within VUELCO. Future unrest crises will allow testing the developed models. As a result the models will not only help the understanding of causes behind volcanic unrest, but will also inform improved scientific and response capabilities to manage volcanic unrest risk.
4. Constraints on pre-eruptive conditions of storage at the target volcanoes create an important framework for interpreting geophysical signals of unrest as well as the resultant mathematical modeling of processes to inform hazard assessment and risk mitigation efforts during unrest crises.
5. The demonstration of significant deviations between experimental simulations of basaltic magma ascent and degassing on one hand and current models for the interpretation of gas data on the other hand, pave the way for future research avenues which may seek to reconcile these discrepancies.
6. We demonstrate that Raman spectroscopy can be performed with a high spatial resolution (1 μm2) and can be used in the field for an approximate chemical analysis.
7. The next-generation densification models developed in VUELCO has important implications for predicting gas budgets and cyclicity at restless volcanoes. The models are applicable in unison with monitoring data to assess if magmatic unrest will likely result in eruption and hance assist hazard and risk assements.
8. Unrest and pre-eruptive geophysical and geochemical data gathered during the project has been made available through open access either from resulting peer-review papers or through the WOVOdat website for interrogation beyond the project.
9. Results from the project led to significantly improved monitoring protocols in observatories run by our partner, the Seismic Research Centre of the University of the West Indies, Trinidad, first of all established in the Montserrat Volcano Observatory, but further rolled out to other observatories in the Caribbean.
10. Similar improvements, particularly in seismic monitoring, have also been introduced at our partner in Ecuador, the Instituto Geofisico, which is responsible for monitoring of all earthquakes and volcanoes throughout Ecuador.
11. The new Bayesian Event Tree has been implemented in an open tool (BET_UNREST) that has been published on the Vhub platform (vhub.org/resources/betunrest ), the reference cyber-infrastructure for volcanological research worldwide. Wider usage of the tool beyond the VUELCO community is expacted as it can be downloaded from Vhub or run online.
12. The BET code calibrated for Campi Flegrei, and used in real-time for the first time in the simulation experiment of the VUELCO project, is currently under consideration by the Italian Civil Protection in a pilot test to make it operational. We foresee that it will become a basic tool to assist decision makers to manage the next volcanic crises in the Campi Flegrei, one of the highest volcanic risk in the world.
13. The Campi Flegrei simulation exercise carried out in the VUELCO project has started a positive interaction between volcanologists and decision makers (the Italian Civil Protection) towards the identification of a future decision making protocol. Such a protocol will allow us to clearly separate roles and responsibilities of each partner involved in the risk reduction process.
14. We have developed guidance notes for stake-holder communication and information and knowledge transfer during unrest crises.
15. The surveys of communities living zones identified as being at high risk from lahars from Cotopaxi have been used by the Instituto Geofisico in Ecuador to alter their local outreach programmes. As a result of the findings they commenced a regular outreach programme to boost levels of confidence of at risk communities in the scientific monitoring of Cotopaxi since June 2013.
16. An awareness of patterns of social unrest is crucial to tailor communication to the lived experiences of at-risk populations. The findings from the Insecurities Methodology demonstrate that pre-crises communication strategies must take into account local sources of insecurity, which has clear implications for volcano observatories and disaster managers. Arguing that a natural hazard should be the most pressing concern when people are faced with real every-day risks (such as crime or unemployment) will likely erode the message and confidence in the disaster management regime.
17. The volcanic unrest simulation exercises developed and execute within VUELCO constitute models that can be usefully exported in other countries exposed to volcanic unrest risk worldwide. The experiences gained from these exercises have informed the preparation of guidance notes. These notes aim at providing a framework to design, run and audit future exercises. Lessons learned during the exercises have informed crises management procedures in the respective jurisdictions and have helped devise strategies to manage unrest crises in the future.
18. Key stakeholders and managers making use of the publically available foregrounds from VUELCO will aid establishing ‘good practices’ regarding the mitigation and governance of volcanic unrest risks. We identified the critical issue of managerial risk vulnerability related to the absence of commonly recognised standards for hazard communications to risk decision makers. This absence may result from the lack of regulation of relevant practices and practitioners. VUELCO was instrumental to help develop checklists for risk governance to represent a rational and methodical way to develop acceptable practice standards focused upon the difficult risk mitigation choices that are made by civil protection authorities and at-risk individuals. The checklist and guidance notes have been adopted and published by the International Association of Volcanology and Chemistry of the Earth’s interior (IAVCEI).
List of Websites:
vuelco.net
Volcanic unrest is a complex multi-hazard phenomenon of volcanism. Uncertainties surrounding cause and outcome of unrest have substantial implications for short-term volcanic hazard assessment. The knowledge of the causative links between subsurface processes, resulting unrest signals and imminent eruption is, today, inadequate to deal effectively with crises of volcanic unrest. This results predominantly from the uncertainties in identifying the causative processes of unrest and as a consequence in forecasting its short-term evolution. However, key for effective unrest risk mitigation and management is the early and reliable identification of changes in the subsurface dynamics of a volcano and their assessment as precursors to an impending eruption.
The VUELCO research consortium has worked over the past 4 years to significantly improve our understanding of the processes behind volcanic unrest and the ability to forecast its outcome aiding decision-making and management in an unrest situation. VUELCO has tackled outstanding problems surrounding volcanic unrest via an international multi-disciplinary consortium, combining fundamental research into causes and effects of volcanic unrest with uncertainty assessment and probabilistic forecasting to improve communication, decision-making and management during volcanic unrest.
The project findings have (i) improved the mechanistic understanding of subsurface processes triggering volcanic unrest, (ii) helped identify reliable precursors associated with specific subsurface processes, (iii) improved the forecasting capacity of the outcome of volcanic unrest in the presence of scientific uncertainty, (iv) improved the capacity for early-warning and management of evolving volcanic crises beginning with monitoring capacity through hazard assessment and risk governance to decision-making and (v) improved the preparedness to cope with consequences and potentially adverse outcomes of volcanic unrest.
Project Context and Objectives:
Volcanic unrest is a complex multi-hazard phenomenon of volcanism. The fact that unrest may, but not necessarily must lead to an imminent eruption contributes significant uncertainty to short-term hazard assessment of volcanic activity world-wide. Although it is reasonable to assume that all eruptions are associated with precursory activity of some sort, the knowledge of the causative links between sub-surface processes, resulting unrest signals and imminent eruption is, today, inadequate to deal effectively with crises of volcanic unrest. This results predominantly from the uncertainties in identifying the causative processes of unrest and as a consequence in forecasting its short-term evolution. However, key for effective risk mitigation and management during unrest is the early and reliable identification of changes in the subsurface dynamics of a volcano and their assessment as precursors to an impending eruption.
When a volcano develops from dormancy through a phase of unrest important scientific, political and social questions need to be addressed. Scientific questions relate to the processes behind unrest and their associated surface signals (if any), their future evolution and their significance as precursor for eruptive phenomena. Political and social questions are related to: (i) how epistemic and aleatory uncertainties surrounding data and processes are employed in the light of probabilistic assessment of outcomes of unrest periods, (ii) how these uncertainties are communicated and (iii) how evolving crises are best managed depending on the forecasted scenario.
The overarching aim of VUELCO is to significantly improve our understanding of the processes behind volcanic unrest and the ability to forecast its outcome aiding decision-making and management in an unrest situation. The working hypothesis behind VUELCO is that current uncertainties surrounding volcanic unrest episodes and can be minimised through a concerted multi-disciplinary and cross-boundary effort.
VUELCO focuses on meeting the objectives via an interdisciplinary consortium drawing from expertise in both Europe and Latin America in three main Project Areas:
1. The research into the mechanistic processes behind unrest and their link to surface observables including capacity building for monitoring and interpretation of unrest signals (Area 1: Science of Volcanic Unrest),
2. The development of tools and protocols for uncertainty assessment and probabilistic forecasting during unrest crises (Area 2: Uncertainty Assessment and Probabilistic Forecasting),
3. The development of risk mitigation programmes, communication protocols and management tools amid scientific uncertainty (Area 3: Communication, Decision-Making and Management).
VUELCO combines fundamental research into causes and effects of volcanic unrest with uncertainty assessment and probabilistic forecasting to improve communication, decision-making and management during volcanic unrest.
The objectives of the project are:
• To improve the mechanistic understanding of subsurface processes triggering volcanic unrest.
• To identify reliable precursors associated with specific subsurface processes.
• To improve the forecasting capacity of the outcome of volcanic unrest periods in the presence of scientific uncertainty.
• To improve the management of evolving volcanic crises beginning with monitoring capacity through hazard assessment and threat analysis to decision-making.
• To improve the preparedness to cope with consequences and potentially adverse outcomes of volcanic unrest.
• To improve the capacity for early-warning of phenomena associated with and the consequences of unrest.
Project Results:
WP3
1. Most volcanic eruptions are preceded by input of less evolved, volatile-rich magma into a shallow-seated magmatic system resulting in magma convection and mixing. Detailed geochemical and petrological investigation at Campi Flegrei, Italy, illustrate several aspects of this relevant process, showing that, although, virtually all eruptions are preceded by magma mixing, not all mixing events lead to an eruption.
2. Numerical simulations of magma convection and mixing in geometrically complex volcanic systems reveal the dynamics of the processes and shed light on the geophysical signals diagnostic of such critical occurrences. Ultra-long-period seismicity is identified as the most indicative signature of on-going magma convection and mixing at depth. Transient events in the frequency band 10-4 – 10-2 Hz, suggested by the numerical simulations, are found from the analysis of monitoring records from strain meters at Campi Flegrei in heightened unrest periods; these are interpreted as the first evidence ever found of magma recharge events at depth during volcanic unrest.
3. 2D and 3D numerical models developed to characterise the influence of the regional stress field on volcanic systems have revealed the relationships between the stress configuration in the lithosphere at different scales (i.e. local, regional and plate-scale) and the distribution of volcanism, particularly in monogenetic volcanic fields. There is a potential for such methods to allow discriminating between volcanic and tectonic controls volcanic unrest. The models developed in the project are of general applicability and can be used to define the stress characteristics of many active volcanic areas, helping interpret unrest episodes.
4. Modelling of magma ascent along volcanic conduits applied to Soufrière Hills volcano on Montserrat (UK), suggests that the observed long-period seismicity can reflect brittle failure of magma at the conduit boundaries. Changes in seismicity patterns might therefore be used as early indicators of changes in ascent dynamics and eruptive styles/impacts.
5. Geodetic modelling at Soufrière Hills reveals that in spite of the current absence of eruptive activity, the eruptive cycle that started in the mid-1990s should not be deemed finished, since the system shows clear signs of over-pressurization. The origin of the over-pressurization is still not clear, and might be due to crystallization induced-degassing (second-boiling) in response to magma cooling. Consequently, the possibility of magma chamber cooling cannot be fully discounted. Extensive computer modelling and new interpretation of the on-going deformation at Montserrat has revealed that a pressurising magma body centred at 6km depth can explain the apparent uplift and lateral deformation. The new results take both topography and the stretching of the tectonic plate into account. This reduces the amount of pressurisation to such a degree, that cooling and crystallisation of the existing magma body is a viable explanation with no or decreased magma influx.
6. A high-precision, two-year long record from White Island, New Zealand was used to test if the movement of volcanic fluids like magma could be detected in a gravity record. While final results of this investigation are still pending, we can conclude that gravity changes in volcanic systems are most likely caused by shallow changes in the hydrothermal system and therefore mask potential magma movement to such a degree that conclusive interpretations of temporal gravity anomalies are very difficult.
7. A novel paradigm for volcano deformation modelling has been developed. The new framework can handle deformation sources of arbitrary geometry that are embedded in realistic rock media characterized by structural heterogeneities, temperature dependent viscoelastic response and complex topography. This approach allows for more accurate estimates of magma reservoirs locations and sizes, providing a substantial improvement in the identification of the possible sources of volcanic unrest. The new generation of forward and inverse models has been successfully applied to Cotopaxi volcano (Ecuador) to identify the magmatic source involved in the unrest phase of 2002-2003; the same modelling has provided insights into the recent eruptive phase (August 2015 onwards). Other applications in the project include unrest deformation in the Altiplano region of Bolivia, Aira caldera, Japan and Soufrière Hills volcano, Montserrat.
8. Numerical simulations of aquifer dynamics and relationships with volcanic activity suggest that the fault system can play a controlling role on heat transport within shallow hydrothermal reservoirs, therefore influencing substantially heat flow measurements at the surface. Results from Campi Flegrei highlight a dominant role of the permeability distribution, with faults representing local maxima in permeability. Simulations of magmatic unrest show that a single magmatic input event can result in an oscillatory response of the hydrothermal system recorded in geochemical (heat flux from the ground) and geophysical (deformation and gravity) signals, as a consequence of complex interactions between gas generation at depth and recharge–discharge dynamics of the reservoir. Investigating the thermal consequences of non-eruptive magma intrusions into glaciated volcanoes as exemplified by Cotopaxi volcano (Ecuador), shows that spatio-temporal variations in rock permeability are the prime contributor to rapid surface heat flux changes, with the potential to produce hazardous ice melt and flooding events without eruptive activity.
9. A comprehensive analysis of decades-long geophysical and geochemical datasets of unrest has been performed at Campi Flegrei in order to explore the capabilities of advanced multi-component gas–melt equilibrium modelling and highly accurate isotopic data to provide the ground for global interpretation of the overall processes subtending volcanic unrest. The results highlight the complexities deriving from the interplay between different degassing sources represented by magma emplaced at different depths below the volcanic system. Such a comprehensive approach demonstrates that complex patterns of geochemical and geophysical quantities recorded at unrest volcanoes can be interpreted in a unique frame, when the competing roles of the different portions of spatially extended, geometrically complex underground magmatic systems are accounted for.
10. Interrogating novel numerical fluid flow models, which simulate fluctuating hydrological recharge in volcanic environments, enabled us to quantify resultant mass redistribution and its effect on geophysical volcano monitoring signals. The results study highlight that mass storage (saturation) changes beneath a survey point, within the unsaturated zone, can generate measurable gravity changes. For a tropical climate inter-annual recharge variations can dominate over seasonal signals and spatially heterogeneous hydro-geological conditions can impact on the accuracy of relative gravity surveys. The additional loading fluctuations associated with saturation variations in the unsaturated zone may also have implications for other geophysical techniques such as monitoring of ground deformation. Ultimately the findings inform the derivation of baseline behaviour of volcanoes in order to better identify pre-eruptive unrest behaviour.
WP4
1. We experimentally determined pre-eruptive processes and magma storage conditions at Tungurahua (Ecuador), Layou & Roseau (Dominica) using a combination of experimental (laboratory-based) and analytical (glass inclusion-based) approaches.
2. We analysed tephra from Mourne aux Diables (Dominica) eruptions in terms of bulk-rock chemistry and mineralogy.
3. We developed an experimental crystal-liquid equilibria database for Cotopaxi Ecuador and applied it to define conditions of the 2015 magmatic unrest at the volcano.
4. We developed mineralogical approaches to constrain pressure-temperature conditions, mechanisms and timescales of basalt magma ascent.
5. We performed experimental simulations of conduit processes (ascent and degassing) for volatile-rich basaltic magmas and found evidence for an anomalous behavior of CO2 and the generation of CO2-supersaturated melts during ascent.
6. We successfully replicated natural glass volatile (H2O, CO2, S) degassing patterns using experiments and showed that gas-melt chemical disequilibrium may occur during magma degassing.
7. We composed a robust compositionally-dependent database for the Raman spectra of natural silicate glasses along the calcalkaline series. In addition, an empirical model based on both the acquired Raman spectra and an ideal mixing equation between basaltic and rhyolitic end-members was constructed.
8. Using magma mixing experiments we developed an efficient mechanism for hybridisation of magma, whereby bubbles originating from a deep mafic melt intruded into a shallow felsic melt.
9. We discovered that the experimentally validated physics of densification in magmas can be divided into two regimes and we derived governing equations for both of them: (a) surface-tension dominated densification, and (b) buoyancy dominated densification.
10. We were able to successfully predict the rates of densification for volcanic systems I support of observations that indicate that magma densification occurs at all volcanoes.
11. We have built a phase diagram that can be interrogated to predict whether or not densification of a system of volcanic particles or bubbles will lead to pressure increases leading to catastrophic fracturing or not.
12. Conditions and processes of phreatic and hydrothermal explosions were investigated experimentally for several targets (Campi Flegrei, Dominica, White Island), increasing our understanding of key aspects of this very variable eruption type.
13. Even small phreatic explosions involving fine-grained ash fall deposits, affected by alunitic alteration, produce a significant and much larger than expected amount of very fines (< 10 µm) with a high potential to instigate chronic diseases.
14. Depending mainly on fluid and rock composition and especially the pH, alteration may both hinder and favour the degassing of hydrothermal fluids and thus the risk of phreatic explosions.
15. Geochemical experiments were successfully tested as a tool to validate reactive transport simulations in volcanic settings. For Montserrat, we were able to fingerprint the response of groundwater to magmatic perturbation.
WP5
1. We have developed and proposed a definition of volcanic unrest: “ The deviation from the background or baseline behaviour of a volcano towards a behaviour which is a cause for concern in the short-term because it might prelude an eruption.”
2. We have performed a global audit of volcanic unrest and created a database from the reported information on unrest at 228 volcanoes. The data is categorised into pre-eruptive or non-eruptive unrest indicators at four different subaerial volcano types and submarine volcanoes as defined by the GVP. Unrest timelines demonstrate how unrest evolved over time and highlight different classes of unrest including reawakening, pulsatory, prolonged, sporadic and intra- eruptive unrest. Statistical tests indicate that pre-eruptive unrest duration was different across different volcano types. 50% of stratovolcanoes erupted after about one month of reported unrest. At large calderas this median average duration of pre-eruptive unrest was about twice as long. At almost five months, shield volcanoes had a significantly longer unrest period before the onset of eruption, compared to both large calderas and stratovolcanoes. At complex volcanoes, eruptive unrest was short lived with only a median average duration of two days. We find that there is only a poor correlation between the length of the inter-eruptive period and unrest duration in the data; statistical significance was only detected for the pair-wise comparison of non-eruptive unrest at large calderas and stratovolcanoes. Results indicate that volcanoes with long periods of quiescence between eruptions will not necessarily undergo prolonged periods of unrest before their next eruption. The resultant database is publically available.
3. We have developed and populated a Dynamic Data Base aimed to process, interpret, and identify seismic precursors that may be causally related to eruptions producing columns of over 4 km height above the crater of Popocatepetl volcano in Mexicos since 1997. The database is available for interrogation and has been incorporated into the WOVOdat master database on volcanic unrest.
4. Multi-parameter unrest data from the El Hierro volcanic crises in 2011/12 has been incorporated into the WOVOdat master database.
WP6
1. An overview was compiled outlining the monitoring techniques already present on each of the target volcanoes. This information was obtained via questionnaires, which were completed by monitoring specialists at each of the volcanoes. This information was then summarised as factsheets for each. This allowed us to identify those techniques with the potential to forecast periods of unrest. It also enabled us to identify the gaps in the existing monitoring network, which might need addressing.
The overview of monitoring techniques already available at the target volcanoes will guide future recommendations for the minimum instruments and processing tools required at each volcano for short-term unrest forecasting.
2. We identified accelerating data sequences within the seismic records and applied the Failure Forecast Method (FFM) as a forecasting tool. The significant result from this is the entire seismic signal must be broken down into the individual waveform components as defined by their cross-correlations. These systems behave according to the FFM and can be used as forecasting candidates. Using the cross-correlation technique in conjunction with FFM provides the most accurate fit of the least squared linear regression of the data.
3. We applied the FFM to an unrest period observed at Chiles and Cerro-Negro volcano, Ecuador. Analysis identified clusters of similar seismic events, but with no change in amplitude through time. This seismicity was interpreted as generated by hydrothermal activity due to the contact of ascending magma with aquifers.
4. By integrating a seismic trace in a routine process, VLPs can be easily identified and later individually processed by removing the instrument response. This allows the detection of ground motion within a frequency band too high, and too small in amplitude for GPS measurements, and too low in frequency to be detectable through usual seismic monitoring procedures. VLPs may play a major role in outgassing events and other sudden movements of volcanic fluids.
5. The numerical integration of seismic traces has now been adopted as a routine processing tool at the MVO on Montserrat and IGENP in Ecuador. Hence, VLP signals can be easily identified and later individually processed and interpreted by removing the instrument response.
6. Through forward modelling, we obtained an entire set of complex volcanic source characteristics in seismic moment tensors that can be used to identify volcanic seismic unrest for different types of seismic signals.
7. We combined tilt and seismic interpretation of Tungurahua activity: Numerical modelling of the monitored tilt signal of Tungurahua volcano, Ecuador, has shown that the signal cannot be explained by the usual inflation/deflation cycles of the volcanic edifice. Pressures necessary to satisfy the data would be unreasonably high. In addition, the fact that explosions on this volcano are often preceded by a ‘deflation’ episode, and therefore depressurization, makes an interpretation difficult. We have demonstrated that an alternative explanation of shear stress induced tilt can explain both the timing of the tilt signal and its amplitude. This will have huge implications for using the tilt signal together with the seismicity as a forecasting tool on Tungurahua and other silicic volcanoes worldwide.
8. SO2, CO2, H2O, HCl and H2S are the most frequently measured gases in volcanic settings and provide a direct insight in magma composition and its temporal evolution. Direct measurements at fumaroles are increasingly replaced by in-plume measurements through ground-based spectroscopic methods in the ultraviolet to infrared spectral range (DOAS, COSPEC, FTIR). We made progress by addressing the weakest point of these measurements concerning the processing step between SO2 snapshot in the plume and calculation of SO2 flux to which a precise determination of wind speed is necessary. We developed an approach using numerical weather models in combination with meteorological data to calibrate local weather stations to estimate the plume transport speed.
WP7
1. We developed a novel prototypal probabilistic forecasting scheme (BET_UNREST) for a Bayesian Event Tree to forecast the probability of occurrence of different hazardous phenomena related to volcanic unrest, both of magmatic and non-magmatic origin. The Bayesian nature of the event-tree permits, on the one hand, to fully account for the uncertainties involved (both of aleatory and epistemic), and on the other hand, to consider all the heterogeneous sources of information: from conceptual and volcanological models to monitoring protocols, the incomplete geological record, and expert opinions. The output of the proposed Bayesian event tree (i.e. the forecast of the probability of different unrest outcomes of magmatic and non-magmatic nature) may support a full management of volcanic crisis, accounting for all the different aspects of volcanic unrest, in particular considering the risk posed by non-magmatic events, such as phreatic eruptions. Such threats are almost always underestimated, but they may pose a major risk during volcanic unrest, as documented by the 2014 Ontake (Japan) phreatic explosion, which claimed several dozen lives. This development represents a key novelty in the field of volcanic hazard assessment, which can aid risk mitigation efforts.
2. The application of Bayesian Event Tree to track the evolution of volcanic unrest at 5 of the 6 target volcanoes of the project, both real-time during the simulation experiments carried out during VUELCO (Colima, Campi Flegrei, Cotopaxi and Dominica), and “offline” for a specific target period (Popocatepetl, April to June 2013). As regards the management of the hazardous phenomena related to non-magmatic unrest, the application of the new Bayesian Event Tree to the Dominica volcanic island, characterized mostly by hydrothermal activity, has allowed the practical usefulness to account for non-magmatic hazardous outcomes in a volcanic unrest crisis.
3. Among the aforementioned applications of the Bayesian Event Tree for managing the volcanic unrest, we highlight the one at Campi Flegrei, as it is one of the highest volcanic risk areas in the world. The event tree provides a real-time estimation of the daily to weekly probability of eruption of different sizes and from different possible vent positions, as a function of the real-time monitoring at the volcano. In a fruitful synergy with another FP7 project (MED-SUV), the results from VUELCO application at Campi Flegrei have been coupled to MED-SUV simulations of tephra fallout, to estimate the short-term tephra fallout hazard, accounting for the wind forecast of the next days. This represents a cutting-edge tool to bridge the scientific knowledge of volcanic unrest and the decision-making purposes.
4. The several experiments carried out during the VUELCO projects highlight clearly the state-of-the-art and strengths and weaknesses of the current interface between volcanologists and emergency managers. This has allowed us to research this link interface in greater detail and we have proposed a general framework based on the hazard/risk separation principle that allows volcanologists and emergency managers to clarify their role and responsibilities in the whole decision-making process.
WP8
1. Using a specifically developed questionnaire, we have performed a global audit of existing communication protocols and practices during volcanic crises. Responses received from stakeholders informed the preparation of best–practice guidelines.
2. The resultant best-practice recommendation and policy-guidelines informed by both the global audit, research conducted in the project and direct interaction with emergency managers and volcano observatory personnel during unrest simulation exercises, aim at developing more efficient and effective inter and intra-stakeholder communication for improved decision-making and management of unrest events.
WP9
1. We audited the level of trust in local authorities and opinion makers around Cotopaxi volcano using questionnaires and field interviews. This set of question became a part of the Insecurities Methodology that has been promulgated to VUELCO partners. General findings include:
• The trust of the Ecuadorian population in authorities is directly affected by broader security concerns. Those authorities which are seen as having a responsibility for every-day security issues, such as crime prevention, are accorded low levels of over-all trust, and this negatively effects the population’s trust in the authorities with respect to a potential volcanic hazard.
• Despite relatively high levels of religiosity, measured with the proxy of church attendance, church leaders are not seen as being particularly trust-worthy in relation to advice on disaster preparedness and response.
• In spaces where scientific expertise is accorded higher levels of trust than government officials, care needs to be taken to ensure that scientists (and their organizations) do not become associated with negative opinions of the broader government.
2. We performed an audit into volcanic risk related decision-making processes in different countries to understand how systems are organized and work. work and are organized. The audit was based on information provided by decision-makers from Mexico, Dominica (W.I.) La Reunion (France), Italy, Ecuador, Canary Island (Spain), Montserrat (UK), and El Salvador.
Key findings included:
• Generally there is a low volcanic risk perception of the general public, even in the large towns close to volcanoes. Decision-makers deem it important that people participate in drills, information campaign, etc.
• Although hazard maps and emergency plans have been generally drawn up (although not regularly updated), exercises and information campaigns have been carried out only in a few countries and for a small amount of people (even in the most exposed populated areas).
• A cost/benefit analysis regarding mitigation efforts is generally not applied.
• Different entities (from Mayor to Ministry) are involved in all levels of communication. Language used is often the official national one only, although sometimes local dialects are also used.
• In order of priority, decision-makers want to know from scientists: 1. kind of expected hazardous phenomena, 2. probability of occurrence of different possible hazardous phenomena, 3. areas with different hazard levels, 4. evolution and duration of the phenomena, 5. time needed to adopt each mitigation action.
• Low importance seems to be given instead to the possibility of false alarms and associated collateral risks. The evaluation of the costs of possible mitigation actions is generally neglected .
• Timeliness of the information appears to be preferred over their accuracy.
3. Prior to the commencement of field work in Ecuador, and before the simulation exercise in Colima, Mexico, a detailed survey of the political and social history of the states was conducted, in particular with relation to security issues. The histories were integrated into the development of the Insecurities Methodology, and informed several of the survey questions. The past history of social unrest in Mexico was also used to inform the audit of the Colima Simulation Exercise. The social history of unrest is a crucial variable in understanding community perceptions of natural hazards and their risk awareness.
4. Social unrest has a direct impact on the perceptions of the state’s disaster preparedness and response mechanisms. Institutions, such as the police, that are seen as a source of insecurity will be accorded lower levels of trust in relation to disaster response. Communication strategies to prepare communities for potential disasters need to take into account the lived experiences of communities. Arguing that a natural hazard should be a central concern in the face of personal and community insecurity will not resonate and may erode the confidence of communities in the disaster management regime.
5. The application of theories of securitization and human security has provided an useful addition to the social vulnerability literature on disaster risk reduction (DRR). It has put into focus the ways in which DRR strategies can increase, rather than decrease, perceptions of insecurity. As the post-simulation audit of the Colima exercise indicates, the state strategy of moving displaced peoples into secure compounds where the freedom of movement is restricted, and family life is disrupted through communal living, if put into practice, is likely to reduce compliance of community members.
6. Significant recent court cases on natural risk governance, including L’Aquila, are framed by process of deduction within a generalised legal infrastructure in order to identify the root causes of the apparent status quo of risk governance. We developed a forensic approach not only to identify the legal responsibilities of societal risk managers and the managerial risks that they face and their causes, but also to consider possible mitigation strategies.
We identify the critical issue of managerial risk vulnerability related to ‘standard equivocality’, which is the absence of commonly recognised standards for hazard communications to risk decision makers. This absence may result from the lack of regulation of relevant practices and practitioners. We offer some recommendations to fuel debate not only within those science groups that reacted to the L’Aquila case but also the scientific community as a whole. Finally, we argue that checklists developed in the framework of VUELCO represent a rational and methodical way to develop acceptable practice standards focussed upon the difficult risk mitigation choices that are made by civil protection authorities and at-risk individuals.
7. We have reviewed the literature on how scientific output from the volcanological community is used in decision-making and we have developed guidelines that may allow scientists and decision makers to bridge the gap between volcanological knowledge and a rational decision making, based on the Cost-Benefit analysis (CBA) of risk mitigation. Basically, CBA is a quantitative study of the trade-offs involved in making a selected mitigating action in the interests of public safety, or of any other stakeholder. As the probability of a specific event exceeds a threshold, defined through CBA, some commensurate mitigation action may become worth taking. This procedure constitutes a crucial link between science and society, establishing rational and transparent decision-making procedures, and clarifying the responsibility of each partner involved in managing a crisis. We have applied the CBA in an illustrative (retrospective) example showing how it could have been used during the VUELCO simulation exercise at Campi Flegrei to minimise the risk posed by pyroclastic density currents. This has been achieved by including results of pyroclastic density currents propagation performed in the FP7 project MED-SUV. This retrospective application to Campi Flegrei can be seen as a first simple and preliminary application.
8. The four volcanic unrest simulation exercises conducted during the project (Colima, Campi Flegrei, Cotopaxi, Dominica) highlighted a number of strengths and weakness including those within: (1) monitoring systems; (2) geological and structural models; (3) local Civil Protection scientific advice systems and emergency management arrangements; (4) communication procedures among stakeholders and; (5) alert level systems.
Based on audits of the simulation exercises we were able was to provide local institutions (scientific, operational and governmental) with important clues for possible improvements in the above-mentioned fields.
The exercises facilitated an extensive assessment of the delicate interactions between scientists and decision makers and the identification of critical relationships and vulnerabilities. They also made possible the testing of probabilistic forecasting models and elicitation methods in simulated “real time” within risk assessment procedures.
The participation of representatives of civil protection authorities from many countries, including those not involved directly in the VUELCO project, was also very important. It encouraged the sharing of expertise, experience and good practices.
In Mexico there was the opportunity to involve the local population in educational activities and a complex evacuation drill, and thereby measure their preparedness and the utility of mitigation measures.
We were able to identify a number of possible enhancements that, if put in place, could really improve the civil protection response to different kinds of emergencies and not only those caused by volcanoes.
Based on audits of the performed simulations, we have devised guidelines to inform the design of future volcanic unrest simulation exercises with global application.
Potential Impact:
Impacts:
In addition to the impact created by dissemination activities of the S&T developments in the project (including conference presentation, direct non-scientific stake-holder interaction, workshops, summer schools, public relations material, press releases, press coverage, virtual agora, peer-reviewed journal publications, video, tweets, peer-to-peer messaging (eg., FACEBOOK)) a number of impacts of the project are specifically highlighted below.
Wider dissemination activities within the volcanological community and beyond included the organisation of many scientific sessions at major international conferences such as EGU, AGU, IAVCEI and IUGG. Scientific papers have been and will continue to be published after the end of funding in high-impact journals.
1. A suite of next-generation physico-mathematical (numerical) models capable of reproducing several aspects of the complex dynamics of volcanic unrest and pre-eruptive processes has been developed and/or refined and improved. Together, these models provide the state-of-the-art reference for the quantitative interpretation of geophysical and geochemical signals recorded at restless volcanoes in terms of magmatic movements, interaction between different magmatic sources, relationships between volcanic and regional/tectonic stress fields, magma degassing, rock properties, presence and dynamics of aquifers. A cook-book to introduce novice modellers to and to validate code from geodetic finite element modelling is available by ‘open access’.
2. Applications of the newly developed models at the target volcanoes and beyond volcanoes highlight the potential of the new modelling resources for new interpretation of the sources behind past and current unrest periods with a potential for a paradigm shift in how to interpret unrest signals.
3. All of the modelling capabilities developed are general and can be applied to any restless volcano, thus, expanding the impacts expected by the mathematical modelling activities within VUELCO. Future unrest crises will allow testing the developed models. As a result the models will not only help the understanding of causes behind volcanic unrest, but will also inform improved scientific and response capabilities to manage volcanic unrest risk.
4. Constraints on pre-eruptive conditions of storage at the target volcanoes create an important framework for interpreting geophysical signals of unrest as well as the resultant mathematical modeling of processes to inform hazard assessment and risk mitigation efforts during unrest crises.
5. The demonstration of significant deviations between experimental simulations of basaltic magma ascent and degassing on one hand and current models for the interpretation of gas data on the other hand, pave the way for future research avenues which may seek to reconcile these discrepancies.
6. We demonstrate that Raman spectroscopy can be performed with a high spatial resolution (1 μm2) and can be used in the field for an approximate chemical analysis.
7. The next-generation densification models developed in VUELCO has important implications for predicting gas budgets and cyclicity at restless volcanoes. The models are applicable in unison with monitoring data to assess if magmatic unrest will likely result in eruption and hance assist hazard and risk assements.
8. Unrest and pre-eruptive geophysical and geochemical data gathered during the project has been made available through open access either from resulting peer-review papers or through the WOVOdat website for interrogation beyond the project.
9. Results from the project led to significantly improved monitoring protocols in observatories run by our partner, the Seismic Research Centre of the University of the West Indies, Trinidad, first of all established in the Montserrat Volcano Observatory, but further rolled out to other observatories in the Caribbean.
10. Similar improvements, particularly in seismic monitoring, have also been introduced at our partner in Ecuador, the Instituto Geofisico, which is responsible for monitoring of all earthquakes and volcanoes throughout Ecuador.
11. The new Bayesian Event Tree has been implemented in an open tool (BET_UNREST) that has been published on the Vhub platform (vhub.org/resources/betunrest ), the reference cyber-infrastructure for volcanological research worldwide. Wider usage of the tool beyond the VUELCO community is expacted as it can be downloaded from Vhub or run online.
12. The BET code calibrated for Campi Flegrei, and used in real-time for the first time in the simulation experiment of the VUELCO project, is currently under consideration by the Italian Civil Protection in a pilot test to make it operational. We foresee that it will become a basic tool to assist decision makers to manage the next volcanic crises in the Campi Flegrei, one of the highest volcanic risk in the world.
13. The Campi Flegrei simulation exercise carried out in the VUELCO project has started a positive interaction between volcanologists and decision makers (the Italian Civil Protection) towards the identification of a future decision making protocol. Such a protocol will allow us to clearly separate roles and responsibilities of each partner involved in the risk reduction process.
14. We have developed guidance notes for stake-holder communication and information and knowledge transfer during unrest crises.
15. The surveys of communities living zones identified as being at high risk from lahars from Cotopaxi have been used by the Instituto Geofisico in Ecuador to alter their local outreach programmes. As a result of the findings they commenced a regular outreach programme to boost levels of confidence of at risk communities in the scientific monitoring of Cotopaxi since June 2013.
16. An awareness of patterns of social unrest is crucial to tailor communication to the lived experiences of at-risk populations. The findings from the Insecurities Methodology demonstrate that pre-crises communication strategies must take into account local sources of insecurity, which has clear implications for volcano observatories and disaster managers. Arguing that a natural hazard should be the most pressing concern when people are faced with real every-day risks (such as crime or unemployment) will likely erode the message and confidence in the disaster management regime.
17. The volcanic unrest simulation exercises developed and execute within VUELCO constitute models that can be usefully exported in other countries exposed to volcanic unrest risk worldwide. The experiences gained from these exercises have informed the preparation of guidance notes. These notes aim at providing a framework to design, run and audit future exercises. Lessons learned during the exercises have informed crises management procedures in the respective jurisdictions and have helped devise strategies to manage unrest crises in the future.
18. Key stakeholders and managers making use of the publically available foregrounds from VUELCO will aid establishing ‘good practices’ regarding the mitigation and governance of volcanic unrest risks. We identified the critical issue of managerial risk vulnerability related to the absence of commonly recognised standards for hazard communications to risk decision makers. This absence may result from the lack of regulation of relevant practices and practitioners. VUELCO was instrumental to help develop checklists for risk governance to represent a rational and methodical way to develop acceptable practice standards focused upon the difficult risk mitigation choices that are made by civil protection authorities and at-risk individuals. The checklist and guidance notes have been adopted and published by the International Association of Volcanology and Chemistry of the Earth’s interior (IAVCEI).
List of Websites:
vuelco.net