Volcanic Crisis Management and Mitigation Strategies: A Multi-Risk Framework Case Study
Natural phenomena such as earthquakes, mud floods, landslides, volcanic eruption or industrial accidents may cause huge disasters. Each crisis generally determines consequences in multiple sectors and systems (e.g. building and infrastructures, roads and public transport, lifelines and telecommunication, society and emergency management.) that are singularly affected by the triggering event, but also may be hit by effects following the main one, the so-called “cascading effects.”
Therefore, the methodological approach to the risk analysis of a single event made by the classical probabilistic convolution of Hazard, Vulnerability and Exposure has to be appropriately modified. Consequently, in order to produce effective tools addressing these issues, dynamic multi-hazard evaluation, time-dependant vulnerability functions of the element at risk under examination, and its exposure evolving in time and space, are required. Systemic studies on hazard, vulnerability and exposure have become the main field of investigation and have to be embedded within a multi-risk assessment approach. Multi-risk in fact is more than a simple aggregation of single risks, due to the possible amplification of risk indexes due to interactions among different hazards and/or cascade events. In this field, only few studies have been conducted and ongoing research still needs considerable efforts.
The tools available for foreseeing crisis evolution are specific to each hazard and therefore the actual best practice for decision makers involved in extreme crisis management is to evaluate possible management sub-plans by comparing potential losses starting from a synoptical view of single risks. Nevertheless, the single risk modeling approaches, based on different hazard models and on the estimation of impacts on typologically different systems at risk, often referring to different space and time scales that determines different model scales and input/output data, may lead to incomparable estimations of risk, ranging from mere indices to complex assessments. Moreover, often only the average of the estimated risk distribution is provided, not accounting for confidence boundaries of the estimation, and at the same time heterogeneity of available input data and models could be accompanied by inhomogeneous uncertainties invalidating risk comparability.
The volcanic hazard represents a relevant research issue in the field of multi-risk assessment and management, as it embodies a possible well-defined sequence of disastrous events (i.e. a Sub-Plinian event is characterized by the following phases: several seismic tremors and earthquakes during the unrest phase, ash fall, bombs and gas emission during the eruption, pyroclastic flows at the end of the eruption and lahars after the eruption), which in turn can lead to cascading effects in multiple systems.
An effective Disaster Operations Management (DOM) tool and its support to decision-makers should then be focused on a dynamic evolution of the risk, in order to minimize the physical and economic losses, especially in the case of Low probability of Occurrence of main event (LO) and High socio economic Impact (HI) of the expected losses. In this case, the effect of long-term mitigation strategies, as well as short-term recovery sub-plans, depending on existing capabilities and on their possible altered availability during the event, should be fully evaluated with a decision management perspective. In this sense a crucial aspect of the dynamic predictive model is the time-dependant vulnerability curves able to estimate the cumulative damage of the considered element at risk by combining the actions of a sequence (seismic crisis) or of the contemporaneity of events (earthquakes with ash on the roofs).
Volcanic Crisis and Disaster Management
Research activities about the effects of a volcanic eruption on existing buildings and infrastructures have produced in the last 15 years a comprehensive framework of studies, surveys and simulations, showing the several uncertainties related to the exact prediction of the eruption type which should be expected (especially in the case of active volcanoes in rest phase), and the many implications that different eruptive phenomena can have on emergency-plan strategies and on built-environment planning and refurbishment.
Nevertheless, in order to address possible DOM (Disaster Operation Management) actions and mitigation strategies in volcanic risk-prone areas, a different approach is needed, starting from the basic consideration that the cumulative effects given by a complex eruptive scenario (such as a Sub-Plinian or Plinian eruption) produce extremely variable impacts on the territory, and depending on the specific time history of the event (i.e. the specific sequence of hazards, from precursory seismic tremors, to ash fall and pyroclastic flows), the prevailing building typologies in the invested areas, and their level of vulnerability (i.e. masonry or R.C. structures).
This peculiar approach has been recently formalized in order to evaluate the impact of a Sub-Plinian eruption in Vesuvius and Campi Flegrei area (Zuccaro et al., 2008, 2009, 2010, Baxter et al. 2008), through the development of a numeric model for the definition of impact scenarios.
The case of a possible volcanic crisis at Vesuvius or Campi Flegrei Caldera can be defined as LO-HD event, since it has very low absolute probability to occur in short time (Low Occurrence – LO), but in that case the impact should be extremely heavy (High Damage – HD). This kind of crisis determines a time-dependent development of the hazard and involves five temporal phases characterized by different types of emergency. The first time-window starts with the unrest phase and is determined when the Vesuvian Observatory of the INGV (National Institute for the Geophysics and Volcanology) records considerable variations of the monitored parameters (seismic activity, geodetic alteration, geochemical variation, etc.). The second phase starts when the eruption happens, with heavy ash-fall, and the third phase consists of possible pyroclastic flows (for large eruption Plinian or Sub Plinian type, Fig.1). Thefourth phase is the post-event period up to the end of the volcanic crisis, characterized by lahars and a possible long waiting period until the monitored parameters come back under prefixed threshold, when the alarm is over. The fifth possible phase concerns the progressive normalization and the coming back to the regular life.
The decisions and activities are then time-dependent and related to hazard phases. As an example, those related to the first time window are basically connected to the population evacuation, while after the main eruptive event, rescue issues also have to be considered in some part of the territory hit by the disastrous phenomena. The fourth phase is very delicate, since it could last years.
It is then clear that many are the scenarios that the decision-makers could be faced with, all in a dynamic time-dependent evolution. The emergency management should be conditioned by several factors. Some of these are not predictable before the eruption starts, such as the intensity of the eruption. Others are potentially predictable in very short time during the unrest phase by using impact scenario models developed in the last 15 years in other UE projects (Vesuvius, EXPORIS etc.) with the contribution of PLINIVS Centre (Università di Napoli Federico II).
Fig. 2 reports a cumulative damage scenario for a Sub Plinian-like eruption at Mount Vesuvius. However, in spite of the prediction capability, many uncertainties arise on how the crisis has to be managed to minimize the impact. In fact, the simulated scenarios are very much influenced by the dynamic modification of the systems along the crisis and this could affect sensibly the actuation of the Civil Protection Management Plan (CPMP) up to his failing.
In order to clarify the level of complexity of disaster-management planning in case of volcanic eruption, it is possible to highlight the main issues associated only with the first phase of the emergency. In fact, during the unrest phase (of which the duration is the first uncertainty) many premonitory tremors and possibly some damaging earthquakes are expected. Therefore, the actual vulnerability assessment of the buildings in the area is supposed to change along the unrest phase by cumulating damages due to the seismic actions that precedes the eruption (Figs 3a and 3b). This could affect the success of the evacuation because of:
• lack of practicability of the road system obstructed by building rubbles;
• emotional pathos of the population disrupted by an emergency (seismic) in another emergency (volcanic);
• management difficulties to rescue the people trapped;
• management difficulties to evacuate relatives of missing people;
• lifeline failures during the management of the seismic crisis preceding the evacuation;
• over-loading of the health facilities in the area and surrounding areas to be evacuated;
• critical management of the information system.
In this case, the decision-makers are faced with the necessity to evaluate the dynamic changes of the system status along the crisis. Therefore, a time-dependent risk assessment is required, along with the evaluation of the variation laws that govern the status along the time history assumed for the several involved systems, such as:
• seismic hazard probability and the dependent eruption probability;
• vulnerability of the buildings facing the escaping ways;
• roads network vulnerability (practicability, redundancies, etc.);
• social preparedness and the probability of bad reactions of the population;
• vulnerability of the information system and management of the media;
• lifelines network vulnerability and its influence on the emergency management;
• rescue system efficiency, time of intervention, capability, etc;
• health system efficiency, time of treatment, capability; etc.
A DOM Tool for Volcanic Events
PLINIVS, as Centre of Competence of the Italian Department of Civil Protection (DPC), has developed a complex model to evaluate damage impact scenarios consequent to a volcanic eruption in Vesuvius and Campi Flegrei area. The tool is based on realistic data for volcanic hazard, vulnerability and losses model that, once translated into possible time-dependent impact scenarios, allows us to simulate the evolution of the crisis, helping decision-makers to choose the possible alternative policies while being aware of the possible consequences in each foreseen evolving scenario. As stated before, the volcanic eruption itself is a sequence of disastrous events, thus determining a significant variation of the vulnerability curves of the element at risk along the eruptive time history. The tool available at PLINIVS simulates the progression of the damage along the crisis, considering the damage variations of the element at risk considered (buildings, people, roads, etc.) in time and space.
The input data of the simulation model are basically:
• seismic activity and his distribution in space, achieved by a sequence of shaking maps representing the damaging seismic event expected during the unrest phase taking into account the geological local effect (Convertito et al. 2008);
• ash fall intensity and spatial distribution according wind speed and direction during the eruption are defined by 3D fluid dynamic simulation model developed at Osservatorio Vesuviano (OV) of INGV (Macedonio et al. 2008);
• dynamic pressure and temperature of the pyroclastic flows and their distribution in time and space at the end of the eruption are defined by 3D fluid dynamic simulation model developed at Centre of Research of INGV in Pisa (Neri et al. 2003);
• potential area invaded by lahars following gas emissions and subsequent heavy rainfall, their dynamic pressure and spatial distribution are defined by overlapping DEM with ash fall distribution obtained by OV simulations;
Inventory data (exposure of the element at risk)
• distribution of building class typologies on the territory;
• total inhabitants and occupants per building type during day/night hours, etc.;
• roads network and critical facilities inventory: lifelines, resources system and their capacity data (hospitals, fire stations, ambulances, helicopters, etc.);
• additional usable resources in “peace time.”
The model can be used as a DOM (Disaster Operation Management) tool, to build up a framework of interactions between the systems involved in the specific crisis considered. The model provides, by means of a full probabilistic approach, the variation of the impact scenario according to the variation during time of the single sub-system, its distribution in space and its influence on the others subsystem interrelated.
Out-of-service road system branches and related no-escape areas are determined. Making use of the hypothetic flow models, traffic intensifications also are determined. Moreover, hypothetical electrical service disruption and gas lines breakings for potential serious inconveniences and/or cascading hazardous effects are determined (systemic vulnerability) involving air transport, trains, etc.
Based on this set of information, the model is able to produce several alternative scenarios at selected times along the event time history, in order to compute an impact evaluation (expressed as casualties, direct physical damages, costs, indirect economic consequences, etc.) to compare the consequences of potential decisions taken (i.e. evacuation of population) during the crisis. The model considers also the probability of variations in damage scenarios consequent to assumed mitigation actions, including the evaluation of the direct, indirect and social cost of each single management action.
The results should be the potential evaluation of a dynamic risk assessment in time and space along the crisis, giving elements to the decision-makers for steering the action to undertake.
Mitigation Strategies and Cost-Benefit Assessment
Studies carried out at PLINIVS, in collaboration with DPC and Campania Region, show how the implementation of mitigation strategies can significantly decrease the expected damage after an eruptive event. Even the impacts of high destructive type of eruptions, such as Sub-Plinian, can be strongly reduced by effective planning strategies or mitigation measures on buildings and infrastructures, responding to the different eruptive phenomena, such as earthquakes (EQ), pyroclastic flows (PF), ash fall (AF) and lahars (LH).
The first domain, related to territorial planning strategies for volcanic risk-prone areas, is connected to some basic regulatory issues that should be faced within local planning and development, introducing the volcanic risk mitigation opportunities into existing or updated building codes.
This raises new issues previously only slightly treated in territorial planning. Incentives to the population to leave the area should be supported by new strategies for the Red Zone, considering how moving many people requires intensive studies for delocalization of residential and productive areas. Some industries and sensible functions would need to be delocalized. But at the same time, having in mind the financial assessment of such operations, it should be clear which kind of use could be envisaged for the new “free” areas in the Red Zone (i.e. urban parks, seasonal and touristic structures, etc.), in order to limit the economic damage to the area due to the loss of people and productive activities. In territorial contexts different from Vesuvius and Campi Flegrei area (characterized by high environmental and archeological constraints to transformation), damage scenarios also can be integrated in advanced territorial planning, suggesting densification and a-densification strategies based on expected impact on the different areas.
A second field of intervention is the one related to the strengthening and protection of building and infrastructures at risk. PLINIVS model considers different mitigation technologies for building structures and envelopes as variable parameters that, if implemented, can reduce vulnerability of technical elements and therefore the expected damage in the area.
In order to define the effectiveness of the technical solutions, the main design data to be considered are:
• construction types and prevalent building technologies in risk-prone areas;
• cumulative effects given by the expected time history of the eruptive event
The design parameters assumed come from the elaboration of data by SPeeD (DPC supported) project (Hazard and Damage Scenarios for Campania Region Volcanoes) and Exploris (EU-FP6) project; the proposed retrofit technologies are directly referred to conventional building types inside Vesuvius and Campi Flegrei area and summarized in a technical solutions database. The technical sheets are classified in four categories: SE-elevation structures; SV-vertical surfaces; SO-horizontal structures; and AP-openings.
The study proposes a design approach for the mitigation of risk, considering the cumulative effects given by the eruptive scenario and highlighting both technical and economical implications, through the definition of performance indicators for the effectiveness of each technical option and a cost-benefit evaluation.
An effective design approach aims to put in relation technological features of existing buildings, parameters and data from probable scenarios, opportunities given by mixing together conventional technologies and advanced materials, highlighting also the additional benefits coming from volcanic risk mitigation such as energy saving issues in case of interventions on building envelope and openings (PF), or value added for householders in case of attic refurbishment overlapping of pitched roof (EQ, AF) with m3 increase.
The analysis of technological options considers performances expressed by the employed materials and technologies, according to some primary requirements (safety, reliability, durability, integrability) and additional selection criteria such as quick installation, storability, lightness, cost, preservation of constructive and architectural features, and multi-functionality (ability to respond to different volcanic phenomena).
In order to transfer this kind of analysis into DOM strategies, it is important to separate “damage connected” mitigation issues (i.e. general strengthening of building structures and roofs, provisional protection systems for windows and openings, etc.) from those that are “emergency management connected” (i.e. seismic mitigation on buildings curtains facing escape routes). With the same perspective, envisaged pre or post-event Civil Protection activities can be supported by “logistic related” mitigation strategies (i.e. protection of railways and train stations as preferred connection system before and after the eruption; protection of strategic buildings in the Red Zone as emergency headquarter, etc.). The envisaged strategies need finally to be supported by cost-benefit analysis.
It is then clear that mitigation strategies must be put in relation with:
• Damage Scenario(s);
• Emergency Management;
• Economic Assessment.
In particular, economic assessment becomes a guideline for future strategies and technical policies. Considering that nobody can predict the duration of a possible volcanic crisis, the cost of the emergency will be very high and strongly dependent from the number of people living in the area.
Scenarios accounting for possible long term mitigation action then have to be considered, such as preventive delocalization of activities, delocalization of entire communities, enlargement of escape roads, population training for emergency situations, “ensafing” established percentages of the buildings through the seismic strengthening of the building facing the escaping roads, or the realization of a pitched roof over the existing flat roof (fig. 4), etc.
The economic assessment tool developed within PLINIVS model enable compares the total financial investment of mitigation measure with the total economic damage after the possible event (Fig. 5).
As an example, a cost-benefit analysis was carried out to evaluate possible mitigation scenarios for ash fall impacts, involving 11 of 59 municipalities in the areas surrounding Vesuvius.
The aim was to “ensafe” about 50% of the buildings through the realization of a pitched roof over the existing flat roof through CFS (Cold Formed Steel) technologies. This solution can significantly reduce the number of victims from a roof collapse, assuming that the people occupying unsafe buildings can find a shelter in buildings subject to mitigation action. The intervention is provided only for the municipalities where the vulnerable roofs areas exceed 50% and the collapsed roof areas exceed 5%. The study has shown that with a total investment of around 182 million Euros, it is possible to reduce of about 35% the number of roof collapsed after the ash fall.
Comparing the expected mitigation cost with the cost of avoided rehabilitation interventions on non-collapsed roof, it is possible to provide a cost-benefit assessment. The results in this case show that the option of such mitigation actions would bring to considerable economic savings in case of a Sub-Plinian eruption of the Vesuvius, varying from about 90.000.000 € (considering an average rehabilitation cost of 500 €/sm) to about 283.000.000 € (average rehab. cost 850 €/sm).
Future Developments and Needs
A dynamic tool that provides the probabilistic evaluation of the consequences of the possible strategies adopted by the decision-makers, before/during/after the crisis (false alarm included), is missing at the moment. Different decisions could affect the damage scenarios consequent to a single eruptive event (that represents a cascade of events by itself). In this aim, the simulation models already available, and mentioned in this paper, consider the cumulative damage on the elements at risk phase after phase (time-dependent vulnerabilities are considered). However, they do not take into account cascading effects that could be triggered during a crisis because of lack of decision or inappropriate decisions.
At the moment, there aren’t models able to describe as a whole the multi-sectorial consequence of an intervention action of the civil protection along the crisis, nor established models able to describe the risk associated to cascading events (multi-risk models) triggered by a wrong managing of an originating adverse event. The decision-maker should be supported by a tool able to represent, through a number of scenarios, what would be the effects of each action. Those scenarios ought to be built considering evolving risk in a multi-risk dynamic approach where the input is also the alternative action of the crisis management.
Future research projects at the European level should take into account this new approach in order to give a first answer to the need of standardization of emergency management, especially when for Low Hazard a High Impact is expected.
• Neri, A., Esposti Ongaro, T., Macedonio, G., Gidaspow, D., 2003. Multiparticle simulation of collapsing volcanic columns and pyroclastic flows. J. Geophys. Res. Lett. 108, 2202.
• Macedonio G., Costa A., Folch A. 2008 Ash fallout scenarios at Vesuvius: numerical simulations and implications for hazard assessment, J. Volcanol. Geotherm. Res.
• Convertito, V. and Zollo, A. 2008. Assessment of pre- and syn-crisis seismic hazard at Vesuvio and Campi Flegrei volcanoes, Campania region southern Italy.
• G. Zuccaro, F. Cacace, R. J.S. Spence, P. J. Baxter, “Impact of explosive eruption scenarios at Vesuvius”, Journal of Volcanology and Geothermal Research 178 (2008) 416–453.
• P.J. Baxter , W.P. Aspinall , A. Neri , G. Zuccaro, R.J.S. Spence, R. Cioni, G. Woo, “Emergency planning and mitigation at Vesuvius: A new evidence-based approach”, Journal of Volcanology and Geothermal Research 178 (2008) 454–473.
• G. Zuccaro, 2009 “A probabilistic Model for the evaluation of the impact of explosive eruption scenarios at Vesuvius”, in Urban Habitat Constructions under Catastrophic Events – COST action C26 – F. Mazzolani et al. Editors. Conferenza COST 26 in Malta 2008 – Pubblicata 2009.
• G. Zuccaro, F. Cacace – 2010, “Seismic impact scenarios in the volcanic areas in Campania” , Proceedings of “COST Action C26- Final Conference” Urban Habitat Constructions under Catastrophic Events – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60685-1
• Nezih Altay, Walter G. Green III, OR/MS research in disaster operations management, European Journal of Operational Research 175 (2006) 475–493
• Kweku-Muata (Noel) Bryson, et al., Using formal MS/OR modeling to support disaster recovery planning, European Journal of Operational Research 141 (2002) 679–688
• Hiroyuki Tamura, et al., Modeling and analysis of decision making problem for mitigating natural disaster risks, European Journal of Operational Research 122 (2000) 461±468
Giulio Zuccaro, born in 1955, architect cum Laude in 1980, Associated Professor of “Scienza delle Costruzioni” at University of Naples “Federico II” in 1999, Director of the PLINIVS Study Centre of the University of Naples, Centre of Competence for the Italian Department of the Civil Protection since 2005. After a professional experience in a multinational engineering company in London (UK), he developed original research on Mechanics of masonry structures, Stochastic Dynamics, Active control of structures, Seismic – Hydro-geologic and Volcanic Vulnerability and Risk assessment. His activity is traceable in more than 120 papers, books and monographic works published in the last 25 years in national and international scientific journals.
Mattia Federico Leone, architect, PhD in Architectural Technology, carries out research activities at the University of Naples Federico II, Department of Urban Design and Planning and Plinivs-LUPT Research Center, in the field of sustainable design, energy retrofit and innovative technologies for buildings and public spaces, with particular reference to advanced materials and technical solutions for the mitigation of natural hazards.