VOLGEO-04278; No of Pages 13

ARTICLE IN PRESS Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j v o l g e o r e s

Improving communication during volcanic crises on small, vulnerable islands W.J. McGuire a,⁎, M.C. Solana a,b, C.R.J Kilburn a, D. Sanderson c a b c

Aon Benfield UCL Hazard Research Centre, Department of Earth Sciences, University College London, Gower Street, London WC1E 6B, UK School of Earth and Environmental Sciences, Burnaby Building, University of Portsmouth, Portsmouth PO1 3QL, UK Centre for Development and Emergency Practice (CENDEP), Oxford Brookes University, Gipsy Lane, Oxford OX3 0BP, UK

a r t i c l e

i n f o

Article history: Received 6 February 2009 Accepted 23 February 2009 Available online xxxx Keywords: communication volcanic crises vulnerability volcanic islands

a b s t r a c t Increased exposure to volcanic hazard, particularly at vulnerable small islands, is driving an urgent and growing need for improved communication between monitoring scientists, emergency managers and the media, in advance of and during volcanic crises. Information gathering exercises undertaken on volcanic islands (Guadeloupe, St. Vincent and Montserrat) in the Lesser Antilles (eastern Caribbean), which have recently experienced – or are currently experiencing – volcanic action, have provided the basis for the compilation and publication of a handbook on Communication During Volcanic Emergencies, aimed at the principal stakeholder groups. The findings of the on-island surveys point up the critical importance of (1) bringing together monitoring scientists, emergency managers, and representatives of the media, well in advance of a volcanic crisis, and (2), ensuring that procedures and protocols are in place that will allow, as far as possible, effective and seamless cooperation and coordination when and if a crisis situation develops. Communication During Volcanic Emergencies is designed to promote and encourage both of these priorities through providing the first source-book addressing working relationships and inter-linkages between the stakeholder groups, and providing examples of good and bad practice. While targeting the volcanic islands of the eastern Caribbean, the source-book and its content are largely generic, and the advice and guidelines contained therein have equal validity in respect of improving communication before and during crises at any volcano, and have application to the communication issue in respect of a range of other geophysical hazards. © 2009 Elsevier B.V. All rights reserved.

1. Introduction A critical problem facing the successful management of volcanic crises arises from difficulties faced in translating expert information from monitoring scientists into appropriate and timely action. The development of an effective warning system depends, to a high degree, on avoiding misunderstanding and confusion in the messages relayed between monitoring scientists, providing the raw data, and those groups responsible for management of the crisis and disseminating warnings to communities at risk, namely the emergency managers and the media. The UN (1995) recognises that a hazard early warning system comprises three elements: (i) the detection of the hazard and assessment of the risk it presents (by the monitoring scientists), (ii) communication of hazard and risk information to those responsible for crisis management (e.g. emergency managers, politicians, government agencies), and (iii) conveyance of warnings to those likely to be affected (the public). For a few hundred of the world's c. 1500 (Simkin and Siebert, 1994) active volcanoes, the first element is reasonably well developed (e.g. WOVO, 2005), the remaining two, however, are often inadequate, or effectively absent. For many, if not most, potentially-active volcanoes, communication between scientists, emergency managers and the ⁎ Corresponding author. Tel.: +44 20 7679 3449; fax: +44 20 7679 2390. E-mail address: [email protected] (W.J. McGuire).

media, and between these groups and the public, is rudimentary at best, and exemplified by the situation that prevailed on the Caribbean island of Montserrat, prior to the start of the continuing crisis brought about by the eruption of the Soufriere Hills volcano in 1995 (Fig. 1) (e.g. Druitt and Kokelaar, 2002). Here, despite, the publication of a report for the Government of Montserrat highlighting the potential risks to the island, and to the capital, Plymouth, of a future eruption (Wadge and Isaacs, 1987), and increased levels of seismicity beneath the volcano in the early 1990s, there was no anticipation of a future eruption and no action taken to put in place measures to handle such a crisis (Sanderson, 1998; Clay et al., 1999). Most crucially, a new ‘disaster action plan’ prepared just two years before the start of the eruption in 1995, in response to the impact of Hurricane Hugo on the island in 1989, failed to acknowledge that a volcanic threat existed (Clay et al., 1999). Even once the eruption had begun, both the Government of Montserrat and the UK Government proved themselves to be unprepared for the communication and public information roles that managing the emergency required (Clay et al., 1999), although this situation did progressively improve, from October 1995, following the first explosive event, and again from August 1997, after lives were lost (Clay et al., 1999). In many ways, the Montserrat volcanic crisis is not typical, in that its drawn-out nature provided a breathing space in which to develop and improve communication between the main stakeholder groups. Fresh magma did not breach the surface until four months after the

0377-0273/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jvolgeores.2009.02.019

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS 2

W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

Fig. 1. Pyroclastic flows from the Soufriere Hills volcano, Montserrat (Lesser Antilles) enter the sea on 17 September 1996, prior to the first explosive eruption later the same day. Courtesy: Bill McGuire.

start of activity, while major pyroclastic flows associated with dome collapse were not evident for another four months. Most volcanic eruptions proceed at a far faster pace, however, with 30% of the c. 3200 eruptions recorded worldwide over within two weeks, and 53% lasting less than two months (Simkin and Siebert, 1994). Consequently, at currently inactive volcanoes where there are no existing communication systems, the chances of being able to put successfully in place such systems, once activity has started, are slim. The most lethal outcome of a serious failure of communication during a volcanic crisis is exemplified by the 1985 Nevado del Ruiz disaster (Voight, 1989), which led to 23,000 deaths (Witham, 2005) – the worst volcano-related disaster since 1902. Simkin et al. (2001) report a significant rise in the number of fatal eruptions over the last seven centuries, which they attribute to global population increase. With an estimated 500 million people (Tilling and Lipman, 1993) or 9% of the global population (Small and Naumann, 2001) at risk from volcanic hazards, and with increasing urbanisation (Chester et al., 2001; Ewert and Harpel, 2004), exposure and vulnerability – particularly in developing countries – making the picture worse, a failure to address communication problems, in respect of volcanic crises, is certain to lead to the loss of many more lives. The communication issue is particularly pertinent to volcanic crises on small islands, where most, if not all, of the population may live within range of potentially lethal hazards, and where the only ‘safe-zone’ may be off island. In such circumstances, organised, rapid, evacuation may be the only response to elevated activity, but is likely to be hindered by limited or inadequate air and sea transport links, and by the population's reluctance to leave the island as opposed to simply their homes. Furthermore, the cost of off-island evacuation may prove prohibitively large, with the result that the government may take a calculated risk to keep its citizens on the island and ‘hope for the best’ – a recipe for potential catastrophe. Other concerns that may act against evacuation include the length of time that the population will remain off-island, and a fear that many citizens will never return. In the case of Montserrat, for example, thousands of citizens have not returned following the evacuation and resettlement of around two-thirds of the population (Clay et al., 1999) between 1995 and 1997. Even if off-island evacuation is not felt to be necessary, it is likely that a significant portion of the population will require evacuation from the vicinity of the volcano to more remote zones of relative safety. This will be accomplished far more speedily and effectively if a communication system has been in place prior to signs

of volcanic unrest, so that the local population is educated with respect to the threat, essential protocols and procedures are known and understood, and the responsibilities of the main stakeholders (those individuals, groups and organisations likely to be involved or impinged upon) are thoroughly ingrained. Given the rapidity with which unrest may escalate to full eruption, the absence of such a system may leave insufficient time to develop one, with potentially lethal consequences. The small sizes of the island volcanoes of the eastern Caribbean, combined with the high level of activity, make them especially vulnerable. Between 1600 and 1899, volcanic activity is estimated (Blong, 1984) to have taken 186,617 lives, while during the 20th century, Witham (2005) has determined a best-estimate death toll total of 91,724. Of the latter figure, fully a third resulted from explosive eruptions on small Caribbean islands – c. 29,000 during the 1902 eruption of Mont Pelée (Martinique), and 1565 at the Soufriere (St Vincent), just a few hours earlier on the same day in May 1902. In addition, the 1976 volcanic crisis on Soufriėre (Guadeloupe) resulted in the temporary evacuation of 73,500 citizens, while most recently, the continuing activity at the Soufriere Hills volcano (Montserrat) took 19 lives in 1997, and led to the long-term displacement of c. 7500 of the islands ~ 12,000 inhabitants. The Lesser Antilles volcanic arc hosts 21 volcanoes that show more or less convincing evidence of activity during the Holocene (Fig. 2) (Lindsay et al., 2005), and seven of which have been active in historic times (post-early 16th century). Discounting the submarine volcano, Kick 'em Jenny, nine eruptions or episodes of minor activity occurred at five volcanoes during the 20th century (Table 1), giving a time-averaged return period of a little over 11 years. 2. Background In light of the worryingly short return period for volcanic activity in the Lesser Antilles, and with failures of communication in the early stages of the Montserrat crisis all too apparent (Sanderson, 1998; Clay et al., 1999), a proposal was submitted to the UK Government's Department for International Development (DFID), with the aim of improving the effective management of forecasting and warning information for small islands vulnerable to volcanic crises. The main objective was the development of a single, integrated, handbook on communication during volcanic emergencies, based upon lessons learnt from recent and continuing volcanic crises in the region. The

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

3

Fig. 2. Active volcanoes of the Lesser Antilles volcanic arc (eastern Caribbean), showing the locations of the volcanic islands of St Vincent, Guadeloupe and Montserrat.

proposed handbook would focus, in particular, on communication between the monitoring scientists, the emergency authorities and the media, both prior to and during volcanic crises, and would be designed as a user-friendly tool for use by all three stakeholder groups. DFID agreed funding and the project started in August 1999, with the

Table 1 Volcanic activity in the Lesser Antilles during the 20th century. Data from Simkin and Siebert 1994 and Lindsay et al., 2005. Economic cost estimates from the University of the West Indies, Seismic Research Centre http://www.uwiseismic.com/General.aspx? id=19. Volcano

Year of Nature of eruption or start of activity eruption or activity

1997 Valley of Desolation (Dominica) Soufriére 1956 (Guadeloupe) 1976 Pelée 1902 (Martinique)

1929

Soufriere Hills (Montserrat) Soufriere (St. Vincent)

1995 1902

1971 1979

Volcano Economic Deaths (Witham, Explosivity cost 2005) Index (VEI) (million US$)

Phreatic

0

?

Phreatic

0

1

0 29,000

2 4

0

3

19

4

N500

1565

4

200

0 0

0 3

100

Phreatic; lahars Dome-forming and explosive; pyroclastic flows; lahars Dome-forming and explosive; pyroclastic flows; lahars Dome-forming; pyroclastic flows Explosive magmatic; pyroclastic flows Dome-forming Dome-forming and phreatomagmatic

1000 1000

handbook completed and available in February 2003. The study focused on the east Caribbean islands of Guadeloupe (Fig. 3). St Vincent (Fig. 4) and Montserrat, where memories of volcanic crises were either still fresh (Guadeloupe 1976; St. Vincent 1971, 1979) or where an emergency was still in place (Montserrat 1995–present). Phase I of the study involved an information gathering exercise, conducted through desk-top surveys, literature reviews and internet searches, focusing on scientific, social and political aspects of volcanic crises at the three islands. Key ‘informants’ were identified and questionnaires drafted for elicitation purposes. Phase II involved field work on the selected islands with the purpose of eliciting responses from key players on their personal experiences with respect to the strengths and weaknesses of communication during past volcanic crises and on their suggestions for improvements and modifications. Elicitation was undertaken through a combination of meetings and semi-structured interviews, with indicators in the following key areas being compared, contrasted and evaluated: forecasting methods; warning messages; warning dissemination; information flow; investment in forecasting and warning; effectiveness of forecasting and warning; organisational culture; coverage of forecast and warning (geography and population); legal background; public education; dissemination of lessons learnt, and; performance targets. Key players from whom information was elicited included staff at both the Guadeloupe and Montserrat Volcano Observatories, representatives of the civil authorities (government, the police and the emergency agencies), NGOs, and the media (press and broadcast), and members of the local population. Analysis of elicited data made possible, for each island, identification of: (1) The strengths and weaknesses endemic in existing communication strategies for volcanic emergencies. (2) Lessons learnt and forgotten from previous crises. (3) The current perception of the volcanic threat.

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS 4

W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

when and if a crisis situation develops. Phase III of the project, involved compilation of its main output – the handbook, Communication during Volcanic Emergencies, designed to promote and encourage both of these priorities through providing the first sourcebook addressing working relationships and inter-linkages between the three, principal stakeholder groups, and providing examples of good and bad practice. We do not view the handbook as a complete and comprehensive solution to the communication issue; we do not argue for it possessing any uniqueness in addressing risk communication in respect of volcanic crises, and we appreciate that adopting the recommendations contained therein is a matter of choice, both of individuals and stakeholder groups. Based as it is, upon the feedback of key players with a strong vested interest in improving communication during volcanic crises, we do, however, hope that the handbook may prove helpful in tackling this difficult and important problem. Phase III of the study also involved extensive review of a draft version, and the collection and analysis of feedback on content, language, presentation and layout, as well as on strategies for the dissemination, usage and further refinement of the handbook. In collaboration with the Caribbean Disaster Emergency Response Agency (CDERA), a draft version of the handbook was presented to the key players on St. Vincent, Guadeloupe and Montserrat, to selected players on other volcanic islands in the Lesser Antilles (St. Lucia and Dominica), and on the non-volcanic island of Antigua, which is exposed to volcanic activity on neighbouring Montserrat. Phase IV focused on preparation and printing of the final version of the handbook, and its dissemination. Copies have been produced as required for distribution to project collaborators, key players on the volcanic islands of the Lesser Antilles, and to appropriate organisations including PAHO (Pan American Health Organisation), IAVCEI Fig. 3. Phreatic eruption at La Soufrière (Guadeloupe) in August 1976. Courtesy: National Geophysical Data Centre, NOAA.

(4) Positive actions that have been shown to be effective on one island and possibilities for exporting these to other islands. (5) Specific complaints among key stakeholder groups about one another, and possible solutions. The key findings of the elicitation process can be summarised as follows: (1) Communication and protocols between emergency managers, monitoring scientists and the media were generally not adequately developed to maximise effective management of future volcanic crises. (2) Structural weaknesses in stakeholder organisations were recognised to be just part of the problem. (3) Low perception of the need to establish effective communications before the development of a crisis was recognised as a more significant issue. (4) The media, as a group, were perceived by both scientists and emergency authorities as – to varying degrees – untrustworthy, and reservations were expressed about its ability to behave responsibly during a crisis situation. (5) Lessons learned during volcanic emergencies as recently as 1976 and 1979 were rapidly forgotten or disregarded. (6) Key players both recognised and welcomed the need for improved communication between stakeholder groups and indicated their enthusiasm for the development of an accessible and usable product that addressed this issue. The findings point up the critical importance of (1) bringing together emergency managers, monitoring scientists, and representatives of the media well before the onset of a volcanic crisis, and (2) ensuring that procedures and protocols are in place that will allow, as far as possible, effective and seamless collaboration and coordination

Fig. 4. A six kilometre eruption column at St Vincent's Soufriere volcano generated during the 1979 volcanic crisis. Photo credit: F.C. Whitemore, USGS.

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

5

(International Association for Volcanology & Chemistry of the Earth's Interior), and WOVO (World Organisation of Volcano Observatories). Professionally produced and packaged E-versions are available on CD, and can be supplied at cost. The handbook is also available in html online, and can be downloaded as a pdf from the Aon Benfield UCL Hazard Research Centre website at: http://www.abuhrc.org/research/volcanic-activity/Pages/project_view.aspx?project=7. A Japanese version of the handbook, translated by Yasuhiro Ishimine of Japan's National Research Institute for Earth Science and Disaster Prevention, can be accessed online at: http://www.bosai.go.jp/library/pub/pdf/caribbean.pdf. The launch of a Spanish version is also planned for 2010.

emergency, however, it is important that monitoring scientists express observations, findings and forecasts in plain language. A trade-off is thus often needed between technical precision and general accessibility. Qualitative, non-technical statements yield more positive reactions among non-scientists – although they must always be based on sound technical analysis. During a crisis, it is important for the scientific team to be clear, to be patient, and to learn from previous mistakes. If a message is not understood, it should be repeated in alternative terms, and as often as necessary, until its meaning has been made clear.

3. Communication during volcanic emergencies

Confusion with terminology is one of the main reasons why nonscientists misunderstand a forecast. Three problems are especially common:

Recommendations made in the handbook are consolidated and presented in the following sections, with respect to the three main stakeholder groups – the monitoring scientists, the emergency managers, and the media. Each section incorporates advice on good practice, on relations with the other stakeholder groups, and on a range of specific, pertinent issues. 4. Monitoring scientists Monitoring scientists are responsible for detecting those geophysical, geodetic and geochemical pre-cursors that might provide warning of a coming eruption, for assessing the probable timing, nature and course of activity, should an eruption be forthcoming, and for providing information on the likely impact of such an eruption. This input is fundamental in relation to determining how emergency managers respond to a crisis situation as it emerges and develops, and it constitutes the first stage in the information chain during a building volcanic emergency. It is thus critical for monitoring scientists to explain their observations, findings and forecasts in plain, unambiguous, language that can be understood by emergency managers, the media and the local population. While scientists may feel tempted, or may be pressured, to issue warnings directly to the public, it is vital that they adhere to a remit that involves the straightforward provision of information to those involved in the management of the emergency. Any initiatives that sidestep or short-circuit this link in the information chain may lead to a loss of credibility as impartial advisors, and unwelcome involvement in the political process. If, on any occasion, a representative of the monitoring team is required to talk to the media, or to make statements or comments related to the crisis, this should be with the full agreement of those involved in the management of the crisis, who should have a dedicated science liaison officer, whose role it is to work closely with the monitoring team. During previous volcanic episodes in the Caribbean, states of crisis have been worsened by misunderstanding of scientific advice and apparent disagreement and conflict amongst the scientists themselves (e.g. Fiske, 1984). This posed a particular problem during the 1976 crisis at Guadeloupe, when up to four independent scientific research teams, were offering contrasting and conflicting advice to the civil authorities (Chester, 1993). To avoid similar problems, an ethical protocol has evolved within the volcanological community to constrain the role and responsibility of scientists during a crisis (IAVCEI Subcommittee for Crisis Protocols, 1999). Although details of the protocol may have to be adjusted on a case-by-case basis, its key features remain unchanged and underpin these guidelines. 4.1. Issuing forecasts Most scientists spend their careers communicating with their peers, and the use of specialised terminology and nomenclature during discussions amongst themselves is natural. In times of volcanic

4.2. Understanding forecasts: problems that non-scientists face when dealing with scientists

(1) The ability to distinguish between terms used in common language that have a different meaning for scientists: especially the difference between hazard (the probability of a natural event occurring), risk (the probability of a loss caused by the hazard) and danger (a situation that can cause damage, injury and/or loss of life). It is important to explain and familiarise the audience with these terms before an emergency arises. (2) A limited public understanding of scientific jargon (e.g., technical terms for volcanic processes, such as “pyroclastic flow”) and concepts such as probabilities in the forecasts. (3) A false understanding of scientific jargon (e.g. Solana et al., 2006). On Montserrat, for example, terms such as “pyroclastic flow” soon entered the public vocabulary, giving the impression that their significance had also been understood. Many times, however, the terms had been learned through repetition by the media, while the true nature and scale of the associated hazards had not been appreciated. Indeed, a false familiarity with the terms may even have lowered the public's perception of danger by inducing an unrealistic sense of security. It is therefore essential for monitoring scientists to explain key technical terms at the start of an emergency (or, if possible, beforehand) and to ensure that non-scientists have properly understood the associated hazard implications. These requirements cannot be understated – a misunderstood message may be more dangerous than no message at all. 4.3. Conveying information effectively With respect to conveying information effectively, most of the elements that make up good practice seem both obvious and selfevident. Nevertheless, many have gone unheeded in previous volcanic emergencies, both in the Caribbean and elsewhere, so need to be highlighted. Messages should be short and simple, and if conveyed verbally through the broadcast media or to the press, should be presented slowly and clearly. If at all possible, the monitoring team should dedicate an experienced individual to all interactions with the media. Only essential information should be imparted. Too much irrelevant scientific data are likely to confuse the audience and obfuscate the basic message. For example; “Earthquakes tell us that…” is more effective than “More than 530 VT earthquakes and near-continuous long period tremors in the last 8 hours mean that …”. Jargon should only be used where essential, and even then should be explained in simple terms. As much as possible, pictures, drawings, graphics, and videofootage should be used to explain concepts, observations and volcanic phenomena. New concepts are best introduced through comparison with an exemplar that is likely to be more familiar with the audience. In this respect, the weather forecast analogy is

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS 6

W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

always a useful one to use when explaining uncertainties in eruption forecasts. Units of measurement that are not in common public use (e.g. Joules, Newtons or Bars) should be avoided, while all numbers, percentages or proportions should be used carefully and sparingly, and backed up by a more general statement. For example; “The probability of Y happening is X %. This means that Y is very likely/unlikely to occur.q Finally, it is vital to confirm that a message has been fully comprehended. To make sure, repeat the message as often as necessary, and in as many guises as possible, until such confirmation is apparent. Bad practice cannot only severely hinder effective communication, it can also build bad feeling towards the monitoring scientists and dangerously sour relationships between the scientists, the emergency managers, the media and the local population. Practices that have been shown to cause problems include: (1) The cultivation of a superior attitude when dealing with groups that are perceived, inevitably, to be less knowledgeable about volcanoes. (2) Condescension to emergency managers. This can rapidly result in the breakdown of working relationships and the potentially fatal collapse of communication and information flow. (3) Adoption of an unnecessarily obtuse or evasive manner. This promotes misunderstanding and leads to fears that something is being hidden. 4.4. The transition from forecast to warning The critical step of transforming a forecast (to advise how a volcano may behave) into a public warning (to advise how a vulnerable community should react) is the responsibility of the authorities. As instigators of the forecast, however, the monitoring scientists are inevitably involved in the transformation process, and have an important role to play in support of the civil authorities. At the height of the continuing activity at the Soufriere Hills volcano on Montserrat, for example, the close link between scientists and emergency managers raised public confidence and improved the response of the population. Most importantly, however, the warnings and instructions were issued by the civil authorities, with the scientists acting only in an advisory capacity. 5. Monitoring scientists and the EMC Scientists provide the information that guides the entire emergency response. This information is then presented to an emergency management committee (EMC), or its equivalent, whose role is to determine the appropriate emergency strategy. From the monitoring scientists' perspective, the clear focus has to be the assessment of the volcano's behaviour. The necessary, frequent, interaction with the EMC can prove a serious distraction and may prove to be counterproductive. In order to minimise this potential problem, it is recommended that a single scientist be given responsibility for liaising with the EMC. Under ideal circumstances, this individual should be part of the committee and work particularly closely with the EMC's science liaison officer, should such a position have been established. This arrangement will allow the core of the scientific team to concentrate on monitoring the volcano. To be most effective, it is imperative that protocols and procedures designed to foster good communication and a strong working relationship between scientists and emergency managers be in place far in advance of a crisis developing. Prior to any sign of activity, it should be decided which scientist will liaise with the EMC or its equivalent. Responsibilities should be agreed with the civil authorities, and the manner and format of information presentation to the EMC should be clarified. Emergency managers should be familiarised with possible eruption scenarios, and aid be offered in

designing basic plans that can be activated rapidly once a crisis starts to develop. The scientists should work with the EMC to develop hazard and risk maps and an appropriate system of alert levels, and offer to take a pro-active role in helping to educate the local population about the nature of the volcanic threat. This is likely to include supplying individuals to talk at public meetings, to schools and other institutions, and on radio and television. Expertise should also be offered in relation to the preparation and dissemination of literature, while scientists should encourage and take a full part in regular exercises and simulations. In order to minimise risk, scientists should also provide prior advice on the safest evacuation routes, making the points that such routes may be affected by precursory activity, including landslides triggered by seismic activity that may block roads and damage bridges, and poor weather, such as the heavy rainfall common in tropical locations. Following the first signs of activity with the potential to build into a crisis situation, priority should be given to ensuring that secure, effective and reliable means of communication are established with the EMC or its equivalent. Bearing in mind that eruptive activity may make travel difficult and damage exposed phone lines, alternative means, such as pagers and VHF radios, should be made available and used. At the earliest opportunity, a timetable for regular meetings with the EMC, or its dedicated scientific liaison officer, should be drawn up. These should take place at least on a daily basis, and more frequently during periods of elevated activity or eruption. The EMC, should be offered the services of the scientists' media spokesperson for press conferences and announcements, for drawing up information releases to the media and the public, and for making presentations to stakeholders such as aid agencies and chambers of commerce. It is imperative that this representative restricts his involvement to explaining the scientific basis for an emergency response and takes no responsibility for emergency management decisions. In order to minimise the potential for a breakdown in the working relationship: (1) Any concerns voiced by the EMC should always be listened to and acted upon as a matter of utmost urgency. (2) Under no circumstances should information be with-held from the EMC. (3) No decisions that might affect the welfare of the local population should be taken without the explicit agreement of the EMC. (4) Anger or frustration with emergency managers should be avoided at all costs; it should always be remembered that they are required to make crucial, life-or-death decisions under conditions of extreme pressure. 6. Monitoring scientists and the media Media attention is commonly a major distraction for the scientific monitoring team, so it is vital that a member of the team is given sole responsibility for liaising with the media. This individual should have good communication skills and previous experience of involvement with the media. Even so, it is likely that individual journalists and broadcasters will still attempt to approach and interview other members of the monitoring team, in part to obtain extra information, but also to find new and different ‘angles’ or to tease out an ‘exclusive’ story. In the latter cases, careless comments can be exaggerated to form the basis of a ‘scoop’, the gist of which is almost always negative and unhelpful, and sometimes dangerous. For example, comments about a population being threatened, even if only theoretically, may be published under the banner headline ‘Population doomed?’, or an apparent lack of consistency about the scientific analysis of the situation may provoke a story with the headline ‘Eruption crisis: what the scientists don't want you to know’. Once such stories have been published, not even later retractions will prevent public doubt about how well a crisis is being managed.

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

6.1. Reducing opportunities for the dissemination of misinformation Appointing a dedicated spokesperson for contact with the media will have the additional benefit of repeat contact with the same scientist encouraging media trust and allowing the core team to concentrate on monitoring the volcano. In order to reduce the chances of media representatives seeking alternative sources or making an independent evaluation about the state of the volcano, the spokesperson should be approachable and media ‘savvy’. To help educate the media, a media pack should be put together, in consultation with the EMC, containing information about the history of the volcano, its activity and its style of eruption. Additionally, the pack should address monitoring methods, contingency plans and appropriate contacts. In collaboration with the EMC, regular, well thought-out and carefully considered press releases about the activity should be issued – even if conditions have not changed since the previous release. Agreement should be reached with the EMC about form and content, prior to making any statements or announcements to the media. Typically, the local media are usually the most effective in informing the population at risk, so giving priority to more glamorous foreign agencies needs to be avoided. It is vital that the same information is released to local and foreign media. Local populations are likely to have friends and relations living abroad with whom they may be in touch during the emergency, and it is imperative that the vulnerable community feel that they are being as well – or better – informed than foreign groups. If the civil authorities have made a specific request that the monitoring scientists should refrain from addressing the media, this should be explained openly to the latter and enquiries redirected to the EMC. If a scientist is required to speak to the media, they should ask, in advance of interview, about the type of information required, so that a general context for the answers can be prepared. In order to minimise opportunities for misinterpretation, the answers themselves should be simple, short, direct and focused. No member of the monitoring team should refuse to reply to a journalist's enquiry without justification or explanation, as this will be interpreted as evasive behaviour that will raise suspicions. Spontaneous, ‘off-the-cuff’ remarks that might be open to misinterpretation should be avoided, while comments should never be made ‘off-therecord’. Above all, the media should never be underestimated or patronised. 7. Communication amongst monitoring scientists It is vitally important that the professional conduct of scientists during a volcanic crisis is impeccable. This is as applicable to relationships between scientists as it is to interaction with emergency managers and the media. Friction between monitoring scientists must be avoided as it will inevitably detract both from the scientific and humanitarian efforts (IAVCEI Subcommittee for Crisis Protocols, 1999). In particular, scientists should never be seen to disagree in public. During the 1976 Guadeloupe crisis, argument and rivalry between research teams on the volcano led to bewilderment amongst the population and confusion amongst the civil authorities (Chester, 1993). Whatever the merits of an argument, open disagreement feeds general distrust and diminishes the authority of the scientific team as a whole. Any contrary points of view should be resolved internally. A number of initiatives can be taken in advance, in order to minimise friction between individuals during a future crisis. Mutually preferred working conditions can be established and protocols agreed for dealing with different opinions within the team, for example adoption of an expert elicitation methodology for forecasting future events during an ongoing crisis (e.g. Aspinall and Woo, 1993; Aspinall and Cooke, 1998). Measures to be enacted once an emergency starts can be rehearsed well in advance, and specific responsibilities allocated. Protocols can be put in place for visiting scientists, defining

7

their tasks and ensuring that they comply with agreed measures for communication with the EMC, the media and the public. For their part, visiting scientists need to be aware that – even if officially invited – they may be regarded as interlopers by some. Potential friction arising from such attitudes can be defused if visitors make a point of deferring to members of the permanent monitoring team and are supportive rather than obstructive at all times. It should be made clear to visiting scientists that they must not engage in tasks or issue statements without the express agreement of the monitoring scientists and the EMC, nor must they make announcements that contradict those of the monitoring team, even if the messages are directed to a foreign audience. Any disagreements need to be resolved internally and in advance. 8. Summarising the roles of the monitoring scientists To optimise its contribution to the successful handling of a crisis, the monitoring team has a responsibility to ensure that relationships with key players and stakeholders are established, and essential protocols and procedures are in place, before the onset of the emergency. These measures should be sufficiently rigorous to operate effectively and with little or no modification as the crisis develops. In summary, they should encompass: (1) Allocation of individuals on the monitoring team to liaise with the Emergency Management Committee, or its equivalent, and the media. (2) Building and fostering strong and supportive working relationships with the EMC and media representatives. (3) Ensuring that the EMC is familiar with possible eruption scenarios. (4) Developing hazard and risk zonation maps based upon the aforementioned scenarios. (5) Working with the EMC and other stakeholders to develop a volcanic emergency plan and ensure that it is regularly and appropriately updated and revised. (6) In advance of the crisis, putting together a checklist of tasks and key personnel to contact. (7) In cooperation with the EMC, planning the means and format of information releases and announcements. (8) Taking a pro-active role in helping to educate the public and the media about the volcanic threat. Should there be no formal relationships between the monitoring team and key players and stakeholders at the start of unrest, the proactive development of such links must form the immediate and urgent priority of the team. 9. The emergency management committee In most, if not all circumstances, the group responsible for managing a volcanic crisis will be the Emergency Management Committee, or its equivalent, whose brief is likely to include the full range of potential technological and natural disasters. Given this broad remit and bearing in mind that volcanic action may not, as at Montserrat in 1995, have occurred within living memory, the first step towards the effective management of a future volcanic crisis must involve the EMC developing a more detailed understanding of the volcanic threat, in the first instance through improved links with the scientific team responsible for volcano monitoring. While every state has its own strategy in place to deal with major emergencies, lessons learned from recent and ongoing volcanic crises in the eastern Caribbean (Guadeloupe 1976; St. Vincent 1971 and 1979; Montserrat 1995–present) can help to improve plans for communication and response. Based to a large extent on the experiences of these islands, this section contains advice on the handling of information during a volcanic emergency and how liaison

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS 8

W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

between the primary stakeholders can be made more effective. It is assumed here that the body responsible for management of the crisis and – thereby – for information handling, will be the EMC. In most Caribbean volcanic island states, this is made up of representatives of key advisory and decision-making bodies, and typically comprises the Prime Minister or Chief Minister, other ministers (as appropriate), the Heads of Police and Civil Defence, a representative of the scientific monitoring team and a spokesperson. This group has ultimate responsibility for decisions impinging upon the welfare of the affected population, including those relating to declaration, changes and cessation of states of alert and/or emergency, evacuation, prohibition of movement and resettlement. Inevitably, an EMC is required to assimilate large amounts of information upon which it is required to act, often rapidly and under considerable pressure. The recommendations that follow are designed to make information handling and communication with other stakeholder groups more effective under these difficult circumstances. 9.1. Investing in community preparedness Community preparedness involves pre-crisis investment in measures designed to minimise the risk of future disaster, and as such forms a critical element in an EMC's planning portfolio. A major complaint of inhabitants living in the danger zones around active and potentially-active volcanoes is that they ‘do not know what is going on’, particularly when the volcano is inactive. The EMC has a key role to play in keeping the local population fully informed before an emergency. This both educates the public with respect to the nature of the threat and improves its response during a crisis. School programmes are a particularly effective way of raising hazard awareness among children and their families, and the EMC should aim to ensure that teaching about the volcanic threat forms a part of school curricula. Other effective awareness-raising initiatives include involving the local population in periodic exercises designed to acquaint the public with procedures and to test and refine emergency measures, and the production and dissemination of educational materials about volcanoes and volcanic hazards. 9.2. Information management during a volcanic crisis Volcanic eruptions in the Caribbean are often not sudden events, and typically the passage from the onset of unrest to the critical stage when they become a serious threat to life, property and general infrastructure, may take months. Such a long build-up should ideally be used by the EMC to activate pre-determined action plans, to ensure that effective links with the scientists and the media are in place, and that the mechanisms for communication of alert levels and warnings are operational. During this build-up phase, the EMC will also have a duty to provide information on the developing crisis to other affected groups such as chambers of commerce, port and airport authorities, airlines, ferries and other transport bodies, and other NGOs. In the months of escalation prior to an eruption, some signals of volcanic unrest (e.g. felt earthquakes or steam outbursts) will, inevitably, become apparent to the population. These are likely to promote some public unease and raise pressure on the EMC to provide more information on the volcano's behaviour. To establish and maintain the public's trust it is imperative that the EMC responds quickly, positively and openly to such requests. Even so, opinion and rumour will feed confusing and conflicting messages, many of which, in the absence of appropriate measures, may be taken up and widely disseminated by the media. Once an eruption starts, there will be little or no time to organise how information is released and communicated or to establish necessary contacts among key stakeholders. It is therefore essential for the EMC to have prepared in advance an effective plan for managing the flow of information during a volcanic crisis.

The names, contact details (telephone, fax and e-mail – at home and at work) and individual responsibilities of the EMC should be known by all committee members. The EMC should meet regularly, even at times of no emergency, to establish close working relations and a shared appreciation of group tasks during a crisis. 9.3. The volcanic crisis information plan Disseminating emergency information is a full-time occupation. During a crisis, the EMC itself will be too busy to pursue this task directly. Through an information liaison officer, the EMC should delegate responsibility for disseminating information to a dedicated information team. Ideally, this would comprise individuals who understand how the government works and are respected and trusted by the public and the media. Candidates with good communication skills are frequently found amongst those with experience as teachers, journalists and media presenters, religious leaders and senior members of government departments. The information team should provide dedicated liaison officers to work closely with the scientists and the media. A ‘cascade’ structure provides for the most effective flow of information. Within this structure, the information team informs a small number of key contacts that in turn pass information to their colleagues and other contacts, with some overlap ensuring that no link in the information flow is broken. In this way, information spreads rapidly, even though the information team spends only a small amount of time alerting their key contacts. Accordingly, a crisis information plan must incorporate a database of the key contacts to alert in case of emergency. This should include reserve names in case the primary contacts are unavailable. The information team should be based at a safe distance from the volcano and within range of the EMC. To avoid unwanted media intrusion at the EMC, it is probably best that the two are lodged in separate buildings, with the location of the information team being public knowledge. 9.4. Disseminating warning messages Warning messages need to be related to a pre-established system of alert levels and should describe both the level of alarm and the required response (e.g. USGS, 2006). Messages must be clear and unambiguous and phrased so as to avoid causing panic or engendering a false sense of security. They should be simple, colloquial, and respectful rather than patronising. It is particularly crucial that warning messages adopt the style of language appropriate to the group being addressed. Use should be made of locally respected leaders, such as mayors, teachers and religious representatives, to spread and reinforce warning messages. Relief agencies may also prove to be useful allies in ensuring effective communication with the public. The message content should be short, simple, and straightforward, include the level of alarm, a description of the expected hazard (s), and instructions about how to respond. The use of pictures and graphics, for example to show evacuation routes, can be highly effective if presented appropriately for the user. Even when there is no change in alert level to be flagged, information should be issued at regular intervals, as a paucity of information may engender unease while frequent updates present an image of awareness, understanding and control. Warning messages need to be consistent, however they are issued, for example whether via radio, the press, disseminated literature, e-mail or dedicated web-sites. Arrangements should also be made so that warnings reach expatriate communities abroad, who are likely to have friends and family under threat. During the first four years of the Soufriere Hills eruption on Montserrat (1995– 1998), the absence of such arrangements presented a particular problem for those Montserratians relocated to the UK, who were largely reliant on informal contacts with friends and relatives still on the island to keep track of the situation (Clay et al., 1999). In order to maximize impact and

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

avoid over-exposure, monitoring scientists and politicians of senior status should be ‘reserved’ for the delivery of warnings that require a major response. Warning messages should avoid difficult concepts and jargon, and should be designed and disseminated so as not to exclude minorities, for example, different versions should target all relevant languages and dialects. Arrangements should also be made to reach those with poor literacy or who are visually or aurally impaired. 10. The EMC and monitoring scientists It is imperative that the EMC creates the conditions that permit the monitoring scientists to focus on their tasks without distraction from other groups, particularly from the media, who may use leaks and incomplete information to create an ‘exclusive’ story. It is equally important, however, that the EMC avoids isolating the scientists from the public, as this can foster a degree of frustration and a feeling that the ‘true facts’ are being with-held as part of a perceived hidden agenda. Well in advance of a crisis developing, the EMC should allocate a science liaison officer from the information team (ideally an individual with some scientific background or knowledge) and ensure that a stable and good relationship is built up with the monitoring scientists. The fact that regular contact between the EMC and the scientific team is being maintained should be publicised, partly to help reduce media pressure on the scientists and partly to strengthen public confidence that preparations for future volcanic emergencies are being taken seriously. Discussion with the scientific team of possible eruption scenarios and their impact, should be used to design a series of basic plans that can be implemented rapidly at the start of an emergency (although these may have to be modified later to account for specific conditions during a crisis). Similarly, hazard and risk zonation maps, developed in collaboration with the monitoring scientists, should form the basis of plans for emergency evacuation and resettlement measures. The sort of information that will be most useful in preparing an emergency plan, should be established with the scientific team, together with the form in which this information will be provided. Prior to a crisis, monitoring scientists will be most amenable and available for involvement in hazard and risk awareness initiatives. Consequently, the EMC needs to use this time to make maximum use of their expertise in helping to educate the local population about the volcanic threat, via public meetings, radio interviews, talks to schools and other institutions, exercises and simulations, and by seeking their involvement in the preparation and dissemination of literature. During a crisis situation, maintenance of effective and secure communication with the scientific team will be paramount. In addition to ensuring that the physical means for such communication are in place, there might also be benefit in basing the EMC science liaison officer at the volcano observatory. A timetable of regular and frequent meetings with the monitoring scientists will also help to maintain the information flow, as will the involvement of the nominated spokesperson of the scientific team in all press conferences, and in preparation of media releases and public announcements. So as to ensure coherence of message, the nature of the information to be released should always be agreed, in advance, with the nominated science spokesperson. Most importantly, the EMC should be responsive to varying circumstances; taking account of scientific advice about the changing behaviour of the volcano and being ready and willing to modify risk maps, the system of alert levels, and evacuation and resettlement plans as and when appropriate. The difference between a well-managed emergency, with no loss of life, and a catastrophe, may lie in the maintenance of a sound and close working relationship between emergency managers and monitoring scientists. To ensure that this relationship is not soured, the EMC should never ignore the advice of the scientific team, nor should it take decisions, without consultation with the scientists, that

9

may impinge upon the welfare of the local population. Members of the monitoring team should not, themselves, be expected or required to produce and/or issue warnings to the public, and should not be pressured into providing forecasts of the timing, style or duration of activity that are beyond current capabilities. Any contacts with external scientists should be conducted through, and with the support of, the monitoring team, in order to avoid friction. Finally, frustration should not be expressed if the scientists cannot guarantee whether or not a period of unrest will culminate in a magmatic eruption. As demonstrated by the 1976 Guadeloupe crisis (Fiske, 1984), this is a notoriously difficult call that, in some circumstances, cannot be made with any degree of certainty. 11. The EMC and the media During a volcanic crisis, the maintenance of good relations between the EMC and the media is vital. Without this, messages can quickly become confused or misrepresented, and considerable effort may need to be expended to ensure that conflicts of interest do not create barriers between the two, which may hinder effective communication and ultimately result in increased risk to the public. Effective collaboration with the media depends to a large extent, upon the EMC – prior to the development of a crisis situation – appreciating how the media operates, understanding the methods it uses to extract and process information, and identifying potential causes of friction. 11.1. The media agenda The ultimate object of any journalist during a crisis situation is to find a ‘good’ story. Typically this will focus upon an ‘angle’ so as to distinguish it from other stories relating to the same event. Investigative journalists, in particular, are instinctively wary and suspicious, and are constantly in search of a ‘cover-up’ that they can address, highlight and try to unravel. They are also notoriously competitive and often attempt to out-do one another. In a crisis situation this can result in increasingly speculative stories based upon ever-more unreliable evidence. In seeking the ‘personal touch’, journalists will consult individuals within the affected population, asking their opinions and extracting information about their experiences. Inevitably, the end-product is often a melange of poorly informed comment and criticism that may reflect badly on the EMC. To validate information, conscientious journalists will cross-check with a number of sources. While increasing accuracy, this also has the potential to highlight differences of opinion and inconsistency of message. A dramatic story is sought after as it captures greater audience interest. Inevitably, something going wrong will be highlighted at the expense of an otherwise flawless operation. Journalists often find it difficult to appreciate levels of scientific (un)certainty, and try and present stories in black and white – thereby highlighting extreme scenarios. In any crisis, they will also seek scapegoats to blame and heroes to praise. In either case this may be entirely unjustified, but it may colour the manner in which the EMC, in particular, and the civil authorities in general, are viewed. Relationships between the EMC and the media may be helped if emergency managers seek to become more aware of how they are viewed by the media. Whether justifiable or not, common media complaints in relation to dealings with emergency managers include suspicions that they are not being kept fully ‘in the picture’, that other journalists are being favoured at their expense, and that the speed of decision-making processes and the paucity and timing of media releases make it difficult to ‘hit’ deadlines. 11.2. Dealing with the media Prior to a crisis developing, the EMC should dedicate a media liaison officer from the information team and ensure that a stable and

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS 10

W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

strong relationship is built up with the local media. Periodic press releases about the state of the volcano – even if nothing is happening – can help maintain links with the media in quieter times. Regular media events, such as visits to the volcano observatory or launches for new monitoring systems or receptions for the arrival of new staff, can all help to build trust and a good working relationship. Advance contact can usefully be made with appropriate journalists (typically science or environmental correspondents) in the regional media and in the major global players such as CNN and the BBC. These are likely to be the first external journalists on the scene when a crisis begins to develop. A database of relevant journalists should be constructed to aid information dissemination during a future crisis, with entries updated as contacts move on and are replaced. Media representatives always appreciate a well put together media pack, which should include information on the history of the volcano, its activity, and its style of eruptions, together with a terminology section. The pack should, in addition, contain information on monitoring, contingency plans for a future eruption, and appropriate contacts. At the start of a crisis, it is recommended – to ensure consistency of message – that a single individual be responsible for all contacts with the media, preferably the aforementioned media liaison officer. All journalists on the database should be contacted immediately, briefed with regard to the situation, made aware of procedures for future dissemination of information, for example the frequency and timing of news conferences and media releases. At an early stage, it is useful to determine which journalists intend to visit and to request details of their travel plans in advance. Suitable facilities should be organised for the media – ideally a room with internet facilities and sufficient e-mail and phone links to allow them to file their stories – and a media pass system organised for journalists who wish to attend news conferences and avail themselves of other official sources of information. This will allow the number and affiliations of visiting journalists to be efficiently tracked and recorded. The media pack should be updated to take account of the changed state of affairs – incorporating information on the volcano's activity, modifications to the monitoring situation, science staff changes, and other relevant circumstances. To ensure that the correct message reaches the media, a regular programme of news conferences and media releases is paramount. The timing of both should be determined at an early stage in the crisis and widely circulated. Under normal circumstances, media releases are best issued at the start and end of the day, with a daily news conference – perhaps during the middle of the day – to allow journalists time to formulate their questions in response to the morning release. Provision of material early in the day is particularly important as it will provide the media with items for their ‘news of the day’ and place journalists under less pressure to hunt for material from unofficial and less informed sources. During periods of elevated activity, after an actual eruption, or following an unusual event – such as a series of notable earth tremors – additional releases and conferences should be arranged. Even if nothing has happened in the preceding 24 hours, the timetabled programme of releases and conferences should be adhered to. With respect to the latter, such circumstances could be used to provide opportunities for addressing specific themes, providing journalists with additional information on certain aspects of the crisis – such as the mechanics of the alert system or the state-of-the-art with regard to eruption forecasting. Members of the monitoring team and other appropriate experts should be utilized at such times. As a matter of course, the monitoring scientists should be consulted in order to make certain that the message the EMC is passing on to the media is concordant with their view. Ideally, a representative from the monitoring team should attend all news conferences, so permitting journalists to acquire first-hand knowledge and reduce suspicion of with-holding of information. Organised visits for registered media representatives to the volcano observatory and into any exclusion zone – when conditions are deemed (relatively)

safe by the monitoring scientists – are useful in that they help to limit unaccompanied, unauthorised, and potentially dangerous, visits. In order to cut short rumour and innuendo, it is important that – as much as is feasible – the media output should be monitored. Inaccurate or malicious reporting should not be ignored, and instead the true situation should be clarified as soon as possible. It may be worthwhile to use news conferences – or specially arranged discussion meetings – to ask journalists for their opinions on how the crisis is being managed and for their impressions about public reaction. Having probably interviewed many individuals across the social range, journalists may be able to provide valuable insights into the public's perception of the crisis and how it is being handled. Above all, bad practice should not be allowed to lead to a deteriorating relationship with the media, so off-the-cuff, off-the-record, and ‘no comment’, responses to requests for information should be outlawed as a matter of course. 12. Summarising the roles of the EMC Effective management of a volcanic crisis depends crucially upon accurate foresight and adequate preparedness and should involve, at the very least, those actions making up the checklist in Table 2. 13. The media The media have a critical role to play during an emergency situation when the unimpeded and effective flow and management of information can literally mean the difference between life and death. In this context, the media's task should be to provide an effective conduit for the transmission of warnings from the EMC to the public, without confusing, complicating, or changing the message. Accurate and responsible reporting is vital if rumour and hearsay are not to lead to unwarranted fear and panic. In the heat of a volcanic crisis, the media's priority should be to fully support emergency managers and monitoring scientists in seeking successfully to handle the situation. 13.1. How others view the media Particularly if no effort has been made to develop a working relationship prior to the crisis, it is likely that emergency managers and monitoring scientists will lack trust in the media and its representatives and have reservations about its reliability as a conduit for information dissemination. Much of this will arise from a general Table 2 Summary of actions that should be taken by an EMC to ensure effective management of a volcanic crisis. • Ensure that a comprehensive disaster management structure forms part of development planning. • Compile a comprehensive checklist of all the steps to be taken as a crisis develops. • Clearly define the duties and responsibilities of key players and groups. • Earmark a single physical base for the management of a future crisis and build an inventory of the equipment, facilities, and services that will be required, including adequate phone lines and other means of communication. • Make sure that a volcanic emergency plan is developed and regularly updated, and made available to all decision makers and stakeholders (e.g. school teachers, local councils, hospitals, chambers of commerce, port and airport authorities, airlines). • Build and foster strong and supportive links with the monitoring scientists and the media. • Update a key contact list regularly and meet frequently and face-to-face with key personnel before a crisis situation arises. • In collaboration with the security authorities, clearly define evacuation protocols, procedures and routes. • Conduct periodic exercises to test and refine emergency procedures. • If the volcanic threat is forecast to impinge upon other states (for example, ash clouds affecting adjacent islands), coordinate information strategies – in advance – with the EMCs of those countries. • Once a crisis has ended, use the experience gained to improve capacity for coping with the next one. Compiling and sharing the lessons learnt, reviewing problems encountered and drawing up solutions for the future.

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

perception of a conflict of interest between the media and other key players in managing a crisis situation, but probably also as a result of previous unsatisfactory contacts with elements of the media. Whether justified or not, journalists are commonly viewed, at least by the public, as being wary and suspicious and always on the look out for a 'good' story, whatever the cost and collateral damage. They are perceived – as a group – as having little regard for accuracy, and for highlighting the theatrical and speculative at the expense of the hard – if less dramatic – facts. Expectations are such that involvement of the media in a crisis situation is widely held to result in antagonism to managers and scientists, with the hunt for stories leading to unwelcome pressures on other key players and the filing of material based upon rumour, innuendo, and unreliable information. It is never going to be easy to overcome this widespread, negative image, but if the media is to play a useful and effective role in the management of a volcanic crisis it needs to work to build trust with other key stakeholders. Ideally, this should begin prior to the development of a crisis situation, although for regional or global players this is unlikely to be a realistic option. 13.2. Good practice in warning dissemination Inevitably, the media will play an important role in warning dissemination during a volcanic crisis – via the press, radio and television and (ideally) the internet. The responsibility of the media can be thought of as being two-fold. First, raising general awareness of the volcanic threat, both prior to and during the crisis, and second, transmission of specific warnings issued by the EMC in response to changes in the behaviour of the volcano and in alert level status. The first point will be of most relevance to the local media, but all media stakeholders are likely to be involved in the latter. Prior to a crisis developing, local press and broadcast media can help raise awareness by working with the EMC and the monitoring scientists to regularly publish or transmit information about the volcano and its status, even when nothing is happening. They can also keep the volcano in the spotlight by writing or broadcasting pieces about the volcano observatory and its work, including new staff arrivals, the applications of new monitoring equipment, open days and public lectures. Similarly, the regular dissemination of information about the alert system, how warnings will be issued at time of crisis, and how the public should respond, would prove useful in terms of maintaining awareness. This is also a good time for the media to (i) decide with the EMC and the monitoring scientists the style that warnings will take, focusing on a simple and straightforward style, the use of pictorial material, and the need to reach all sections of the community, (ii) perfect effective and rapid lines of communication with the EMC and monitoring scientists – via their media liaison representatives – to be utilized at time of crisis, and (iii) develop and update web pages focusing on the volcano and its activity, the warning systems, and how information will be disseminated during a crisis. In a crisis situation, the local media should ensure communication links with the EMC are maintained via its media liaison officer and with the monitoring scientists' spokesperson. Sufficient airtime and column space should be allocated to incorporate new information about the developing crisis, with provision made to break into scheduled programmes for urgent announcements. In order to attract listeners' attention, broadcast media may consider adopting an immediately recognizable jingle or ‘intro’ for all announcements related to the crisis. In the press, an instantly recognisable logo can be used to attract attention in a similar way, linked to news about the volcano that always occupies the same space – preferably on the front page. Web pages developed by local media at the pre-crisis stage need to be updated, while the regional and global media might be persuaded to build new ones from scratch. Of particular importance is consistency of message, and this should be maintained both off the

11

island as well as on, and – for multi-media organisations – should be uniform across all channels. Bad practices may sour relations with the EMC and the monitoring scientists, and may ultimately result in lives lost. In this regard, the following should be avoided: (1) Issuance of any announcements related to the volcano's behaviour or to alert levels or warnings without the express agreement of the EMC. (2) Embellishment of messages with unnecessary information. (3) Broadcasting or printing of information or opinion from unofficial sources without prior consultation with the EMC. (4) Failing to issue warning messages immediately they are received. 14. The media and the EMC During a volcanic crisis, the EMC – or its equivalent – will have one priority, to limit injury, loss of life, and damage to property and infrastructure. All else will be secondary. Once a crisis begins to develop the EMC and its individual members will be under extreme pressure. To avoid being intrusive and distracting, media representatives should restrict day-to-day contact with the EMC to the media liaison officer. As appropriate, he or she may be able to arrange interviews with other members of the EMC. Media representatives should be aware that many politicians and senior administrators – who are likely to be key players during the emergency, and some of whom will undoubtedly sit on the EMC – will have misgivings about liaising with the media. Typical concerns voiced by those involved in emergency management include the following: (1) Answers will be misinterpreted or twisted. (2) Vital parts of the message they are trying to present will be omitted, modified or confused. (3) The message will be made incoherent and inconsistent. (4) False information will be released to the public. (5) The situation will be over-dramatised. 14.1. Establishing effective working relations with the EMC Overcoming the aforementioned preconceptions is an essential pre-requisite for developing an effective working relationship with the EMC. For the local media, this can be accomplished by liaison before a crisis develops. At this time, the material in media packs should be digested and used, and a good relationship built with the EMC media liaison officer. The local media should endeavor to work with the EMC in advance of a crisis so as to design a plan for the dissemination of warning messages, and to clearly define the role of the media in this critical aspect of emergency management. Regional and global players arriving on the scene once a crisis has already started to develop need to inform the EMC of their arrival and proceed via formal channels and with the knowledge and agreement of the EMC and its media liaison officer. All media representatives should make a point of attending official news conferences and joining organised events such as visits to the monitoring observatory or into exclusions zones. Bad practices that are likely to lead to friction in relationships with the EMC are to be avoided, and include: (1) Arriving on the scene unannounced. (2) Shunning contact with the EMC and its media liaison officer. (3) Short-circuiting official routes of information. This can sow bad feeling and mistrust and confuse or damage warning messages. It may also distract scientists and emergency managers from their primary tasks. (4) Looking for hidden agendas that are not there or inventing scapegoats that may hinder the effectiveness of crisis management.

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS 12

W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

(5) Inventing, encouraging or spreading, rumour or innuendo. Unofficial views should always be cross-checked with official sources. 15. The media and the monitoring scientists In an emergency situation, members of the monitoring team will be fully occupied and under considerable pressure. Consequently most will rarely be able to spare the time for lengthy interviews with the media. It is likely, however, that – in order to limit interference with the work of the team as a whole – an individual will have been given the role of liaising with the media. 15.1. Establishing effective working relations with the monitoring scientists In order not to interfere with the monitoring effort, all approaches for information and comment should be made through the monitoring team's media spokesperson. Bearing in mind that they must first clear statements and announcements with the EMC, they should not be pressurised for comment in advance of this. To speed up the procedure, the gist of a proposed interview should, wherever possible, be provided in advance. Verity and accuracy in reporting a scientist's replies can benefit from the interviewer being familiar – in advance – with the local volcano, its behaviour and history, so as to be able to ask sensible and valid questions, and comprehend the replies. If an interviewer does not understand something that a scientist has said or written, they should ask for clarification. Misinterpretation could have serious implications for management of the crisis. A strong, working relationship will be helped if media representatives are straightforward and open with the scientists, rather than seeking to trick them into revealing information that could harm management of the emergency. The science team will be hindered by harassment, by attempts at pursuing disagreement or fomenting dissent amongst the scientific community, by seeking ‘off-the-record’ comment, by deliberately misinterpreting, modifying or embellishing the message provided by a member of the monitoring team, and by promoting the views of scientists who are not team members and who may hold maverick views. 16. Relations within the media The jockeying for competitive advantage that often characterizes interaction between media players has no part in an emergency situation. Impeding the functioning of rivals, attempting to with-hold significant information and spreading misleading messages to put other media players off 'the scent', may compromise the effective dissemination of warnings and thereby increase the threat to the local population. A possible problem lies in the potential for contradictory information to be issued by different media sources, leading to confusion, doubt and even fear amongst the public. To minimize this all media players should focus on the core message as presented to them via the Emergency Management Committee, without embellishment, modification or interpretation. 17. Summarising the roles of the media To an important degree, the media are in a position to make the difference between an effectively managed volcanic crisis and a catastrophe in which many lives are lost. Ensuring the former rather than the latter requires that the local media, in particular, take those actions summarised in Table 3. Notwithstanding these actions, it has to be recognised that the role of the media is inevitably a complex one. While perhaps wishing to be supportive, journalists will also rightly feel that they have a duty to investigate and report how effectively the situation is being handled,

Table 3 Summary of actions that should be taken by the media to ensure effective management of a volcanic crisis. • Become familiar with the volcano in question and the community or communities likely to be affected. • Ensure that trustworthy working relationships are developed with the EMC and the monitoring scientists before a crisis develops, or as soon as feasible during its early stages. • Use all available means (press, television and radio and the internet) pro-actively to promote the work of the monitoring team and the EMC contingency plans amongst the affected population. • Together with the EMC and the monitoring team, decide on the form and style of the warnings to be issued during a crisis situation. • Make certain that provision is made for effective and robust lines of communication with the EMC science liaison officer and monitoring team spokesperson. • Devise protocols to ensure consistency of message. • Ensure that pride of place – in terms of both airtime and column space – is devoted to information about the crisis and associated warnings. • Formulate and use an instantly recognisable ‘brand image’ for crisis information dissemination (a logo or ‘theme’ tune). • Promote vigorously, the warnings provided by the EMC, without embellishment, modification or dramatisation. • Avoid the short-circuiting of official routes of information. • Work with the emergency managers and the monitoring team – be a help not a hindrance.

and should be free to hold the EMC to account if there is poor or ineffective management. 18. Discussion Population rise and increasing urbanisation are driving a marked upward trend in the incidence of natural disasters. This trend is already being reflected in the numbers of lethal volcanic events (Simkin et al., 2001), and in a rise in the number of volcano-related deaths. Tilling (1990) estimates that the average number of deaths per year due to volcanic activity, in the 17th–19th centuries, was 315, while Witham (2005) shows that this figure has climbed to 917 for the 20th century. Witham (2005) reports 491 volcanic ‘events’ during the 20th century, in which people were killed, injured or affected in some way, and more than half of which resulted in loss of life, while Chester et al. (2001) and Ewert and Harpel (2004) draw attention to increased exposure and vulnerability arising from the continued growth of urban centres close to active volcanoes. At least 500 million people currently live within the danger zones of active volcanoes, and this number is certain to climb substantially. Despite the threat they pose, volcanoes have always attracted settlement due to characteristically fertile soils. With global average temperatures set to rise significantly this century and beyond (IPCC, 2001), volcanoes located in the tropics will also provide increasingly appealing havens for those seeking the cooler conditions brought by higher altitude. Elsewhere, and particularly in the Caribbean, attractive volcanic islands will bring in increasing numbers of tourists. The pattern of volcano-seismic crises at Montserrat's Soufriere Hills volcano suggests an elevated risk of another eruption (assuming a prior cessation in current activity) around 2025 (Clay et al., 1999), while seismic swarms beneath the island of Dominica between 1998 and 2000 – along with a small phreatic explosion in 1997 – have led to a probabilistic estimate of a magmatic eruption as high as 1 in 5 between 2000 and 2010 (University of the West Indies, Seismic Research Unit, 2000). Certainly, the time-averaged frequency of eruptions and minor activity during the last century, of a little over 11 years, demands immediate action on the communication issue, rather than once a future episode of unrest has begun. 19. Conclusions As populations rise on the flanks of active and potentially active volcanoes, so procedures and protocols to deal with future volcanic

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019

ARTICLE IN PRESS W.J. McGuire et al. / Journal of Volcanology and Geothermal Research xxx (2009) xxx–xxx

crisis must be better thought out and increasingly stringent. Most importantly, emergency plans must address the issue of seamless communication between the main stakeholder groups, and ensure that the mechanisms that underpin effective communication during a volcanic crisis are in place long before a volcano shows signs of unrest. This is particularly critical for small, volcanic islands, where the options for managing a volcanic crisis are more limited than for mainland volcanoes. For the island volcanoes of the Caribbean, where eruptions are frequent, such measures need to be in place now or in the very near future. Messages from all the stakeholder groups must be as simple and concise as possible and must address instructions and recommendations on what to do next as well as inform of the prevailing situation. All efforts need to focus on the building of trust between stakeholders, the maintenance of good working relationships, and the safeguarding of an open and continuous information flow between all key players. We hope that the advice and recommendations presented in this paper and in the handbook for Communication during Volcanic Emergencies, will help improve the management of volcanic crises in the Caribbean and on other small volcanic islands. As the source-book and its content are largely generic, we also feel that the suggestions and guidelines contained therein are equally valid in terms of improving communication before and during crises at any volcano, and have application to the communication issue in respect of a range of other geophysical hazards. Acknowledgments This research was funded by the UK Department for International Development; Project No. R7406 Protecting vulnerable small islands by improved forecasting and warning. The following are gratefully acknowledged for their invaluable contributions to the project: Nichola Brichieri-Colombi, Myrium Lubino-Bissante, Charles van Oppen, Jeremy Collymore, Noemi D'Ozouville, Eviann Inniss and Olly Willetts. David Chester and Angus Duncan are thanked for their helpful and insightful suggestions during the review process, which have greatly improved the paper. References Aspinall, W.P., Woo, G., 1993. A formalised decision-making procedure for assessment of volcanic eruption threats using expert judgement. Proceedings SECED/Royal Society IDNDR meeting. Protecting Vulnerable Communities. London. 13 pp. Aspinall, W.P., Cooke, R.M., 1998. Expert judgement and the Montserrat volcano eruption. In: Mosleh, A., Bari, R.A. (Eds.), Proceedings of the 4th International Conference on Probabilistic Safety Assessment and Management PSAM4. 13 – 18 September 1998, New York, vol. 3, pp. 2113–2118.

13

Blong, R.J., 1984. Volcanic Hazards. Academic Press, London. 424pp. Chester, D.K., 1993. Volcanoes and Society. Edward Arnold, London. 351pp. Chester, D.K., Degg, M., Duncan, A.M., Guest, J.E., 2001. The increasing exposure of cities to the effects of volcanic eruptions: a global survey. Environmental Hazards 2, 89–103. Clay, E., Barrow, C., Benson, C., Dempster, J., Kokelaar, P., Pillai, N., Seaman, J., 1999. An evaluation of HMG's response to the Montserrat volcanic emergency. UK Department of International Development Evaluation Report EV635. London. Two volumes (86 and 182pp). Druitt, T.H., Kokelaar, B.P. (Eds.), 2002. The eruption of Soufriere Hills volcano, Montserrat, from 1995 – 1999. Memoir, vol. 21. Geological Society, London. 645 pp. Ewert, J.W., Harpel, C.J., 2004. In harm's way: population and volcanic risk. Geotimes 14–17 (May). Fiske, R., 1984. Volcanologists, journalists and the concerned public: a tale of two crises in the eastern Caribbean. Explosive Volcanism: Inception, Evolution and Hazards. National Academy Press, Washington DC, p. 179-176. IAVCEI Subcommittee for Crisis Protocols, 1999. Professional conduct of scientists during volcanic crises. Bulletin of Volcanology 60, 323–334. IPCC, 2001. Third Assessment Report: Climate Change 2001: The Scientific Basis. Cambridge University Press. 881 pp. Volcanic hazards atlas of the Lesser Antilles. In: Lindsay, J.M., Robertson, R.E.A., Shepherd, J.B., Ali, S. (Eds.), Seismic Research Unit. University of the West Indies. 279 pp plus maps. Sanderson, D., 1998. Volcanic warning dissemination in Montserrat. In: Lee, B., Davis, I. (Eds.), Forecasts and warnings. UK National Coordination Committee for the International Decade for Natural Disaster Reduction. Thomas Telford, London, pp. 5.34–5.41. Simkin, T., Siebert, L., 1994. Volcanoes of the World. Geoscience Press, Tucson. 368 pp. Simkin, T., Siebert, L., Blong, R., 2001. Volcano fatalites – lessons from the historical record. Science 291, 255. Small, C., Naumann, T., 2001. The global distribution of human population and recent volcanism. Environmental Hazards 3, 93–109. Solana, M.C., Nave, A., Kilburn, C.R.J., Sorensen, S.A., 2006. Hazard awareness among decision makers at volcanoes in repose: theory and practice. Cities on Volcanoes 4 Abstracts, 23–27 January 2006, Quito, Ecuador. Tilling, R.I., 1990. Reducing volcanic risk: are we winning the battle but losing the war? Earthquakes and Volcanoes 22, 133–137. Tilling, R.I., Lipman, P.W., 1993. Lessons in reducing volcanic risk. Nature 364, 277–280. United Nations, 1995. Early-warning capacities of the United Nations system with regard to natural disasters. Secretary General Report A/50/526. 28pp. University of the West Indies, Seismic Research Unit, 2000. Volcanic Hazard Report for Southern Dominica: Interpretation of 1998–2000 Earthquakes and Hazard Mapping Results. 49pp. http://www.uwiseismic.com/Documents/REPORTS/ Dominica/Final%20Report.pdf. United States Geological Survey, 2006. USGS Alert-Level System for Communicating the Status of Activity at US Volcanoes. http://volcanoes.usgs.gov/2006/vhpalertlevel.pdf. Voight, B., 1989. The 1985 Nevado del Ruiz volcano catastrophe: anatomy and retrospection. Journal of Volcanology and Geothermal Research 42, 151–188. Wadge, G., Isaacs, M.C., 1987. Volcanic hazards from Soufriere Hills volcano, Montserrat, West Indies. A report to the Government of Montserrat and the Pan Caribbean Disaster Preparedness and Prevention Project. Department of Geography. University of Reading, UK. Witham, C.S., 2005. Volcanic disasters and incidents: a new database. Journal of Volcanology and Geothermal Research 148, 191–233. World Organisation of Volcano Observatories, 2005. Directory of Volcano Observatories 2005. http://www.wovo.org/dir-contents.htm.

Please cite this article as: McGuire, W.J., et al., Improving communication during volcanic crises on small, vulnerable islands, Journal of Volcanology and Geothermal Research (2009), doi:10.1016/j.jvolgeores.2009.02.019