CHAPTER 12

Enabling effective hazard management by the public

COLIN GREEN

 

Introduction

With no sense of irony, we frequently congratulate ourselves upon the heroic behaviour of the emergency services in a disaster. If the emergency services have to be heroic, then hazard management has failed since the objective is to reduce a potential disaster to the easily manageable. Equally, the emergency services frequently have to compensate with heroism for the failures of command, communication and control during the disaster, and mismanagement and negligence prior to the disaster.

Moreover, focusing attention on the emergency service response concentrates both upon a particular form of response – that by the emergency services – and intervention at a particular point in the disaster process – the aftermath of disaster. Instead attention should focus on the objectives of hazard management and the assessment of the relative effectiveness of different intervention strategies.

By focusing upon a particular type of response at a particular stage, as an overall strategy, hazard management fragments into different unrelated strategies and policies where emergency planning becomes an isolated activity and an end in itself, lacking clear criteria to judge the effectiveness of alternative strategies. Emergency planning becomes what emergency planners and the emergency services do and tends to focus upon those hazards for which the emergency services have the clearest role. Thus, the best prepared emergency plans are probably those for hospitals in terms of dealing with mass casualties from a disaster (Lee 1989). This is a rather defeatist approach to emergency planning and is mirrored by a neglect of plans to warn and, if

necessary, evacuate people in a danger zone. It also results in a tendency to plan to manage things rather than to manage people; the public typically are treated as objects rather than as independent decision-makers.

Context

We cannot, of course, eliminate hazards and hazard management must be set within the context of overall social, economic and environmental management. It is within the latter constraints that the objective of hazard management of minimising the likelihood and consequences of a disaster is set. We have to live with hazards. The question for hazard management is: how great a threat must be tolerated given other objectives and the available strategies for reducing the threat. Or, to use current terminology, what is the ‘tolerable risk’ (Health and Safety Executive 1988). The threat or risk which is tolerable from a given hazard is strictly defined as that which results from the least worst alternative amongst those available (Green 1988a).

This chapter is deliberately entitled ‘hazard management’, as opposed to ‘emergency planning’, because I would like to distinguish between ‘hazard management’, ‘the management of emergencies’ and ‘disaster management’. Figure 12.1 illustrates the way that hazard management includes both management of an emergency and disaster management as time-specific subtasks of hazard management. Since a particular hazard does not necessarily give rise to an emergency, while hazard management is a continuing necessity, the latter two functions are ones which hopefully will not occur. Indeed, the success of hazard management is measured by its success in reducing the probability of the necessity of managing an emergency, and the success of emergency management by the degree to which the need for and task of disaster management is reduced. Self-evidently, hazard management is a process and I want also to emphasise that ‘planning’ is simply the preparation and rehearsal for a situation which will subsequently have to be managed. Planning is therefore only valuable in so far as it results in more successful management.

Hazard management is a process over time which starts with the identification or recognition of the hazard as presenting a potential threat (Figure 12.1). This begs the questions of who recognises and how we recognise a potential hazard: too often these only occur because a disaster has occurred. Therefore, an essential part of a hazard management policy is a strategy for identifying potential hazards so that they can be managed.

Following Haddon (1973), a hazard can be defined in terms of the potential of an unplanned energy exchange, since the mechanism which causes damage is an energy exchange. ‘Hazard management’ is preferred as a term to the more conventional ‘risk management’ because in reality we do not manage risks but hazards. If we consider the existing level of risk from a nuclear power station design to be intolerable, then we may add extra containment to the reactor vessel or add diversity or redundancy to the operating and safety systems. In consequence of these changes the level of risk will be reduced. Similarly, the risk of flooding is reduced as a consequence of widening or deepening the river channel, building retention reservoirs, or building flood embankments. ‘Risk’ is an abstract concept which summarises some of the consequences of our decisions.

Hazard management involves both the assessment of the hazard plus the selection of the appropriate intervention strategy. Hazard assessment is a comparison of the potential magnitudes of energy release – hazard analysis – and their probability of occurrence versus the vulnerability of any exposed populations – vulnerability analysis. Clearly if there are no people or vulnerable installations, then any unplanned energy exchange from the hazard will have no effect. Thus, the magnitude of the threat posed by the hazard is always a function of the potential magnitude of the release and the vulnerability of the exposed population and installations. Consequently, potentially the threat can be reduced by decreasing the vulnerability of the exposed population.

Since the process of hazard management is a sequence over time, it makes logical sense to start disaster management at the beginning rather than at the end. The need to plan to manage disasters is a sign that hazard management has been ineffective.

Although I have emphasised that by its nature hazard management is a unified process, in practice the management and planning of the latter two stages may be separated administratively from the hazard assessment and intervention stages. This is not a call for some unified agency to take responsibility for the whole of hazard management for each of all hazards. This might be the optimal solution for managing the hazard in question, but in terms of management or government of society as a whole, it is likely to be sub-optimal. The process cannot be arbitrarily split into three stages: co-ordination and liaison between the lead agencies involved in each stage are essential, and equally coordination and liaison are necessary between the multiplicity of organisations involved within each stage. Hazard management at each stage is about co-operation and co-ordination: the ‘I'm in charge’ syndrome usually results in inefficient management.

It is also unhelpful to define a disaster in terms of the outcome, the degree to which successively hazards, emergency and disaster management have all failed. Instead we require a definition which indicates the nature and scale of the management task for which we have to plan. In management terms a ‘major incident’ can be distinguished from a ‘crisis’ in two ways (Green and Parker 1985) (see also Table 1.1 above). In major incidents the ratio of available emergency personnel to the affected public can approach or exceed 1 to 1 whereas in crises the ratio will be very much less than 1 to 1. A second differentiating factor is the length of time available between the precursor event and the onset of the energy exchange. Events where this is short to non-existent are less challenging because there is little or no possibility of warning those at risk or doing anything further to modify the energy exchange. The two different events pose radically different management tasks. Major incidents involve co-ordinating highly trained and disciplined emergency services. Successful management of a crisis event must rely upon channelling and facilitating the response by the public.

Objectives of hazard management

The objective of hazard management is to reduce the likelihood of a hazardous event occurring and/or to reduce the consequences should one occur. Lazarus's (1966) model of stress can be usefully generalised to describe the stress which potentially hazardous events pose to society. In this model, stress is a function of the challenge or demand from the event versus the resources available to meet that challenge. Hazard management is therefore successful to the degree to which it reduces the likelihood of a challenge; or reduces the magnitude of the challenge; or mobilises sufficient resources to overcome the challenge.

We can then redefine vulnerability to a hazard as the likelihood that a given individual, household or society will fail to have sufficient resources mobilised to meet the challenge posed by that hazard. For example, the resources which an individual can mobilise are dependent upon the physical and financial resources available to them, their knowledge, the degree of social support available to them, their health and personality, and often what can be variously described as their social competence. Social competence refers to the extent to which individuals can fill out the forms necessary to claim insurance; whether or not they avoid being browbeaten by the insurance assessor (Green et al 1985); have political power (Green 1989); are willing to complain; or can organise effectively so that someone pays attention.

If successful, hazard management can reduce the likelihood or magnitude of the challenge. Equally, unsuccessful hazard management can magnify the challenge or reduce the resources available to meet the challenge. For example, in most British floods the residents can sit out a flood by retreating upstairs. However, it has been noted elsewhere that the best way to map floodplains in Britain could well be to map the locations of single-storey sheltered housing for the elderly and mobile home parks (Green and Parker 1985). Consequently, a common response to the risk of future flooding by householders living in a bungalow is to install a loft ladder (Green et al 1985).

The vulnerability of a household to floods can be expressed in the form of Figure 12.2. The characteristics of the flood in terms of depth, duration, type of flood and so forth define the challenge; and the characteristics of the dwelling either magnify or reduce this challenge, as well as providing resources to be mobilised to meet the challenge. The characteristics of the household are measures of the resources which are available to be mobilised to meet this challenge (Green and Penning-Rowsell 1985). In turn, this model can be elaborated to represent the different impacts of a flood, the relationships between these impacts and the resources available to mitigate each impact (Green et al 1987). Subsequently, those who are in poor prior health are more likely to experience stress during a flood event and more likely to experience long-term health damage (Green 1989). Not surprisingly, households with greater financial resources report a lesser effect from a given financial loss than do households with few financial resources (Green et al 1985).

The resources available to the individual need to be broadly drawn; personality characteristics such as coping style (Wheaton 1985) have been hypothesed to affect the stress an individual will experience from an event of a given magnitude. So similarly has the social support available to the individual or household both from family and friends as well as other groups such as the social services (Thoits 1982). There is evidence that social support can mitigate the long-term effects of stress. Those who report experiencing a high level of stress during a flood are more likely to receive social support mitigating the effects of the stress experienced (Green 1989).

However, social support, which is often spontaneous, is frequently poorly targeted both in terms of who is supported and the form of support provided. Lions Clubs and similar groups often take away flooded carpets to be dried.

The result can be a shrunk carpet covered in hard mud. A form of help which is more appreciated is assistance in filling forms, arranging for contractors to repair property and generally helping in those areas where the volunteer from the Lions Club has greater social competence or power than the flood victim. In the 1978 Wisbech flood, one of the main functions of the social services was persuading victims' insurance companies to pay the claims (Fearns 1978). Flood victims' experience of insurance companies is variable.

During a crisis one of the reasons why communication rates are so high, and may even completely saturate the telecommunications system, is that people respond to the perceived challenge by mobilising social support. Thus, in the 1984 flood at Uphill, daughters telephoned mothers, wives telephoned husbands and mothers telephoned their children for help before the flood cut off the telephones (Green et al 1985).

In terms of life safety, we can similarly develop a model of vulnerability in the fault tree for the risk of death from flooding (Figure 12.3). The transition probabilities are a function of the challenge (depth of flooding) and resources (structural stability of the building, prior health).

Public hazard management

Management of a ‘crisis’ necessarily must be based upon enabling the public to adapt effectively to the threat. Hazard management which attempts to ignore the public or treat the public as passive will fail. The job of government is to use its scarce resources, together with information, to help the public adopt the most effective response, or to improve the effectiveness of the response adopted by the public. The largest part of rescue and action during the emergency period is carried out by the victims (Dynes et al 1987). This is confirmed by the television films of both the Hillsborough disaster and the Bradford fire disaster. Earthquake planning in California is quite explicitly based upon the recognition that the victims will have to cope on their own without outside assistance for 48 hours. Equally, the influx of volunteers to a crisis area is also both a typical and a documented feature of disasters (Wolensky 1979).

All hazard management strategies, including emergency and disaster plans, contain assumptions as to behavioural responses. Off-site emergency plans for industrial sites identified under the Control of Industrial Major Accident Hazards (CIMAH) Regulations are based on the assumption that in the event of gas release, people will go indoors and close the doors and windows (Health and Safety Executive 1989). Similarly, off-site emergency plans for nuclear power stations assume that nearby residents will go indoors, and wait delivery by the police of iodide tablets or orders for evacuation. The realism of these assumptions, either that the public will know what to do or that they will wait until they are told what to do and then do it, must be questioned. For other hazards, notably fire, hazard management strategies contained similar assumptions as to behavioural responses (Canter 1983, 1990) which when examined proved either to be impractical (e.g. use of fire extinguishers) or inaccurate (e.g. the use of exits from buildings).

In an emergency there is no reason to expect people to stop making decisions using available information and acting accordingly. Nor is it realistic to assume that people will wait until someone tells them what to do. The opposite occurs if people perceive a clear threat to their own or others' safety, not least because the autonomic physiological response to such a threat prepares them for ‘fight or flight’. Similarly, advice they receive about evacuation and shelter will be interpreted in the light of their prior expectations and integrated with other information.

People's beliefs about a hazard determine their expectations about the behaviour that will be appropriate should a crisis occur (Green 1980). Consequently, the public are only likely to take shelter rather than attempt to leave, if their expectations of the hazard lead them to decide that sheltering is the most adaptive strategy. The autonomic response for ‘fight or flight’ means that to take shelter is to fight physiology; therefore, people will have both to be particularly convinced of the appropriateness of the response to adapt it (Green 1990) and capable of coping with physiological stress resulting from its adaptation.

The generally low-key nature of information given to residents about the nature of the risk and the actions they should take in the event of an emergency means that the public's prior expectations will be crucial. Some leaflets (e.g. Sellafield Local Liaison Committee 1982), consist of four pages of reassurance that the event could never happen before receiving behavioural advice on how to respond. This may be contrasted with the graphical and pictorial type of information provided in other countries and is probably unlikely to change people's likely behaviour.

Nor can it be assumed that people will either instantly identify a cue or signal as a warning. All signals are ambiguous and the most probable interpretation is seldom that the signal constitutes a warning. For example, the ratio of non-fire warnings to fire warnings from fire alarms is about 10:1. Early research on human behaviour in fires showed that people sought confirmation of a fire before responding to an alarm (Canter et al 1980), thereby eroding the warning lead- time. The response has been to institutionalise fire drills to shape the public's behavioural response to evacuate. The role of fire wardens is essentially not to check the building in a fire but to force compliance in drills.

Consequently, if a government is to use its superior information to enable effective public adaptation to hazards, then knowledge of the public's behavioural expectations is essential. Secondly, the government must deter- mine the most effective points of intervention. Figure 12.4 is a process model of hazard management which indicates points where intervention in the public's management of the hazard is possible (Green, Parker and Penning-Rowsell 1988).

Cultural factors

In practice, cultural factors can quite clearly influence the effectiveness of hazard management. Britain is relatively free of extreme climatic and geotechnic hazards. The term ‘natural hazard’ is deliberately avoided here because it is so very misleading. ‘Natural hazards’ is a very broad descriptive term, and hazards described as natural have no more in common than do graphite, coal and diamonds which are all forms of carbon. The dichotomy between ‘natural’ and ‘man-made’ is false, and can divert attention away from the real question of choosing the best strategy to manage a particular hazard.

Probably because of this absence of climatic and geotechnic hazards, emergency planning in Britain became identified with civil defence for war. Traditionally, from World War II onwards, the primary function of civil defence in wartime in Britain has been to preserve the State. Apart from major civil unrest, few other hazards present the same risk of survival to the State as does war and thus they have been a low priority on the political agenda.

We also have an institutionalised view of the public as emotional, irrational and ignorant: if the public knew the risks, they would only panic. Consequently, they shouldn't be told anything. In the Second World War, the government was very concerned that a collapse of civilian morale would undermine Britain's ability to continue the war. Morale was seen as an emotional state which could be boosted by the appropriate propaganda. Since it was therefore necessary to determine the state of civilian morale, much effort went into developing a way of measuring morale, not least so that the effectiveness of propaganda aimed at improving morale could be determined. Eventually, the best measure proved to be the question: ‘How confident are you in the government's ability to win the war?’ Not only is this a fundamentally rational approach but confidence, or morale, also improved or declined according to the progress of the war – with a decline after a disaster

(e.g. at Singapore) and an improvement after a victory (e.g. at Stalingrad) (McLaine 1979).

In spite of the persuasive evidence that people do not panic in a disaster except in very extreme circumstances, the concept of panic is pervasive as a label for behaviours. What seems to be happening is that we label people as having panicked when they adopt behaviours which do not seem to us afterwards as having been those which maximised their own or others' survival chances. That is, in perfect hindsight knowledge of the consequences of the behavioural options open to the victims, those who made the wrong choice are said to have panicked.

In developing a scale of the stress imposed by a flood upon individuals, one of the statements used was ‘I panicked’. This was one of the few items which distinguished between males and females: male respondents were significantly more likely to agree with the statement than were women (Green 1988b). I tend to interpret this finding as indicating that males believe that they ought to know what to do in a situation, and regard themselves as having panicked when either they didn't know what to do or in hindsight they believe that they did not adopt the most effective action. Panic therefore seems to be little more than the label attached to behaviours which fail due to inadequate information at the time the decision had to be made as to the behaviours possible or the likely success of each behaviour. More usually, the problem is to persuade people that the situation is sufficiently critical that they should act (Leik et al 1981).

Planning for management

Planning should be a preparation for the problems which management will confront. The problems which the public and professional hazard managers face in managing an emergency are largely identical: unreliable information, acute time pressures, an inherently improbable event and inadequate communications.

Thus, in the case of chemical hazards, it will often be that sheltering is the most adaptive strategy (Prugh 1985): it also is a behaviour which should take little time to implement if adopted. If the elapsed times for the evacuations at Mississauga (Liverman and Wilson 1981) and Barking (Stanbridge 1980) are compared to times for toxic cloud dispersal (Gray 1981) and toxicity (Prugh 1985) at different distances out from a source, then to be effective, the process of initiating a call for evacuation must be taken at a time when the likelihood of a release occurring is not at all certain and well before the release has occurred.

To these problems is added for the hazard manager the need to liaise and co-ordinate with a multiplicity of different groups. Consequently, planning should be seen as a rehearsal for the event. The planning process should be seen as being more important than the product, and as a way of rehearsing the liaison and co-operation between organisations which will be essential to effective management.

Suggestions for improvements in British hazard management

The following are a number of suggestions for improving hazard management in Britain.

  (i) Emergency plans are necessarily based upon assumptions or expectations about human behaviour which must be validated. Emergency plans must either be based upon the public's behavioural expectations, or attempts must be made to influence the public's behavioural expectations towards more adaptive behaviour.

 (ii) Instructions on preparing emergency plans are all too often invocations: e.g. develop a warning system. It is necessary to be able to assess the likelihood that emergency planning will be effective without running a live experiment, and to develop hierarchies of criteria against which alternative options can be compared.

(iii) It is desirable to apply the principles of reliability analysis (Kaufmann 1972) to emergency planning as well as in hazard management (Gruetter and Schnitter 1982). This analysis can be used to assess the likely effectiveness of a plan and to identify those components where improvements would be most cost-effective.

 (iv) A key component of integrated hazard management is a form of technological forecasting: the identification of hazards which are likely to emerge. Hazard management needs to be pro-active rather than reactive; we need to stop bolting the stable door after the horse has bolted, or, more bluntly, after we have recognised the hazard because there is a pile of bodies.

TABLE 12.1 Dam failure

Dams illustrate the Law of Large Numbers: that if there are enough examples of dams, then however low the probability of a dam failure, the probability that one dam will fail is high. In Britain there are 537 large dams (International Commission on Large Dams 1983, Bossman-Aggrey et al 1987). These are mainly earth embankment dams. The most likely cause of failure is a passive mode failure. There is no publically available information on the risk of such failure, although the Safety and Reliability Directorate of the United Kingdom Atomic Energy Authority has carried out a reliability analysis (Parr and Cullen 1988). But the lower bound on the risk can be set by the comment from an author of that study that ‘if the nuclear industry applied to operate a plant which posed the same level of risk as an earth dam they would never be granted an operating licence’ (Tyrer 1988). Since the Health and Safety Executive sets a target of 1 × 10⁻⁶ per annum per plant for the risk from a nuclear plant, it can, therefore, be assumed that the risk of failure by an earth embankment dam exceeds this figure.

If the risk of failure were as low as 1 × 10⁻⁶ per year per dam, then the chance of a dam failure in any year would be 1 in 1862. If the risk of failure were as high as 1 × 10⁻⁴ per year per dam, close to the world average figure, then the chance per year of a dam failure somewhere within Britain, would be 1 in 19.

In the case of a dam emergency, the likely scenario is that the dam keeper is likely to identify the precursor event. For earth embankment dams, Owen (1980) suggests that the time to failure is likely to be between 0.5 and 4 hours. The dam keeper will probably be a manual-grade employee, and consequently is likely to phone his/her office seeking instructions; no decision to advise of an evacuation is likely until a qualified engineer has visited the site. An hour or more could well elapse before a qualified engineer reaches the site and decides that an evacuation should take place.

The nearest police station could well be manned only in office hours, but once the Headquarters has been informed, and decided that an evacuation is justified, again with some elapse of time, the police will then be faced with the problems of deciding who needs to evacuate, how to warn them, where they should evacuate to, and how they should get there. At the present time in Britian, there are no maps identifying the area which will be affected by the flood wave following a dam failure although the preparation of such maps is required in some other countries (Dunglas 1988, Owen 1982, Province of British Columbia 1984). At night the only way to warn people will be to go ‘on the knocker’. Police will therefore have to be mobilised and dispatched to the area. In the absence of information on which residents require assistance to evacuate, the police will have to identify these people when they are warned and then call up transport. The daytime situation would be better in some ways in that local radio could be used to communicate warnings, but with the complication that members of family groups who are separated are likely to seek other family members before they will leave.

Consequently, I am not nearly so sanguine as Parr and Cullen (1988) in anticipating that evacuation will be successful in the time available. Instead, I expect casualties to be high.

been much cheaper to deal with the problem of methane release from waste disposal sites If houses had not been constructed upon them before the risks were identified. The aim of management is to achieve a surprise-free future; currently there is a tendency to prepare for the last disaster rather than to try to plan for the one which has not happened yet.

If I had to predict disaster which will occur in Britain, then probable events Include:

  (i) a major dam failure;

 (ii) a train crash involving hazardous chemicals;

(iii) a repeat of the 1947 Thames flood between Oxford arid Teddington this time resulting in the evacuation of about 30,000 people for up to three weeks;

(iv) a ‘Love Canal’ situation – a housing area found to have very high levels of contamination from heavy metals and other hazardous metals: most probably on the site of an old gasworks;

 (v) a contamination of flood waters by chemical waste; and

(vi) the collapse of a sea wall followed by major urban flooding (Figure 12.5).

This is by no means a complete list, and a major aircraft crash on an urban area is deliberately omitted because this is everybody's favourite disaster exercise.

  (v) The institutionalisation of learning both from actual incidents and from experience with other hazards is essential. It is tragic that when counselling services were being established for victims of the 1984 Bradford football stadium fire, the standard North American material was not available (Lystad 1985, National Institute of Mental Health 1979).

 (vi) Preparation and maintenance of an emergency plan should be seen as a rehearsal for managing an emergency and not as a document. The process which involves establishing communication, liaison and co-operative linkages replicates the problems of managing the emergency. The actual document is less important than the construction of the collaborative network which has gone into its preparation.

(vii) Emergencies are always low probability events and consequently both individuals and organisations will always have difficulty in adapting to them and, indeed, in recognising them. Regular rehearsals are the only way of increasing the likelihood that both individuals and organisations will both identify the need to respond and respond appropriately.

REFERENCES

Bossman-Aggrey, P., Green, C.H. and Parker, D.J. (1987). Dam Safety Management in the United Kingdom, Geography and Planning Paper No. 21, School of Geography and Planning, Middlesex Polytechnic, Enfield.

Canter, D.V. (1983). Studies of Human Behaviour in Fire, Department of Psychology, University of Surrey, Guildford.

Canter, D.V. (1990). Fires and Human Behaviour (2nd Edn), Fulton, London.

Canter, D.V., Breaux, J. and Sime, J. (1980). Domestic, Multiple Occupancy, and Hospital Fires, in Canter, D.V. (ed.) Fires and Human Behaviour, 117–136, John Wiley, Chichester.

Drabek, T.E. (1986). An Inventory of Sociological Findings – Human System Responses to Disaster, Springer-Verlag, Berlin.

Dunglas, M.J. (1988). La Securité des barrages, Bulletin No. 22, Conseil Général du Génie Rural des Eaux et des Forêts, Paris.

Dynes, R.R., de Marchi, B. and Pelanda, C. (eds) (1987). Sociology of Disasters: contribution of sociology to disaster research, Franco Angeli, Milan.

Fearns, R. (1978). The response of Cambridgeshire Social Services Department to the Wisbech flood of January 11, 1978 – an Assessment of the Operation and the Lessons for Future Planning, Unpublished M.A. Thesis, University of Essex.

Gray, J. (1981). Characteristic Patterns of and Variations in Community Response to Acute Chemical Emergencies, Journal of Hazardous Materials 4, 357–365.

Green, C.H. (1980). Risk: Beliefs and Attitudes in Canter, D.V. (ed.) Fires and Human Behaviour, 277–291, John Wiley, Chichester.

Green, C.H. (1988a). The Tolerability of Risk from Nuclear Power Stations: A critique, Report prepared for Friends of the Earth, Flood Hazard Research Centre, Enfield.

Green, C.H. (1988b). The relationship between flooding, stress and health, paper given at the London Meeting of the British Psychology Society.

Green, C.H. (1989). Stress and Social Support, paper given to the British Disaster Research Group.

Green, C.H. (1990). Perceived risk: past, present and future conditional in Handmer, J. and Penning-Rowsell, E.C. (eds) Hazards and the communication of risk, 31–52, Gower, Aldershot.

Green, C.H., Emery, P.J., Penning-Rowsell, E.C. and Parker, D.J. (1985). The health effects of flooding: a survey at Uphill, Avon, Flood Hazard Research Centre, Enfield.

Green, C.H. and Parker, D.J. (1985). Civil Disaster Management: an academic perspective, paper given at the Civil Defence College, Easing-wold.

Green, C.H., Parker, D.J. and Penning-Rowsell, E.C. (1988). Flood Hazard Management by Public and Government, paper given at the 1st Meeting of the European Chapter of the Society for Risk Analysis, Laxenburg.

Green, C.H. and Penning-Rowsell, E.C. (1985). The evaluation of the intangible benefits and costs of a flood alleviation proposal, Journal of the Institution of Water Engineers and Scientists 40 (3), 229–248.

Green, C.H., Tunstall, S.M., Emery, P.J. and Bossman-Aggrey, P. (1987). Evaluating the non-monetary impacts of flooding, paper given at the Annual Meeting of the Society for Risk Analysis, Houston.

Gruetter, F. and Schnitter, N.J. (1982). Analytical risk assessment for dams, paper given at the ICOLD 14th Congress on Large Dams, Rio de Janeiro.

Haddon, W.D. (1973). Energy damage and the ten countermeasure strategies, Journal of Trauma 13 (4), 321–331.

Health and Safety Executive (1988). The tolerability of risk from nuclear power stations, HMSO, London.

Health and Safety Executive (1989). Risk criteria for land-use planning in the vicinity of major industrial hazards, HMSO, London.

Herschy, R., Newsome, D. and Green, C.H. (1989). The Benefit-Cost Hydrometric Data – River Flow Gauging, Report to the Department of the Environment, CNS Scientific and Engineering Services, Reading.

International Commission on Large Dams (ICOLD) (1983). World Register of Dams, ICOLD, Paris.

Kaufmann, A. (1972). Reliability: a mathematical approach, Transworld, London.

Lazarus, R.S. (1966). Psychological Stress and the Coping Process, Academic Press, New York.

Lee, I. (1989). When an aeroplane catches fire: the Manchester International Airport disaster, 1985 in Walsh, M. (ed.) Disasters – Current Planning and Experience, Edward Arnold, London.

Leik, R.K., Carter, T.M., Clark, J.P., Kendall, S.D., Gifford, G.A. and Ekker, K. (1981). Community Response to Natural Hazard Warnings, University of Minnesota, Minnesota.

Liverman, D. and Wilson, J. (1981). The Mississauga Train Derailment and Evacuation, The Canadian Geographer 25, 365–375.

Lystad, M. (ed.) (1985). Innovations in Mental Health Services to Disaster Victims, National Institute of Mental Health, Rockville.

McLaine, L. (1979). Ministry of Morale: Home Front Morale and the Ministry of Information in World War II, Allen and Unwin, London.

National Institute of Mental Health (1979). Crisis Intervention Programs for Disaster Victims in smaller communities, US Department of Health and Human Services, Washington.

Neal, J. and Parker, D.J. (1989). Flood warnings in the Severn Trent Water Authority area: an investigation of standards of service, effectiveness and customer satisfaction, Middlesex Polytechnic, Flood Hazard Research Centre, Enfield.

Owen, H.J. (1980). Flood Emergency Plans: Guidelines for Corps Dams, Hydrologic Engineering Center, US Army Corps of Engineers, Davis, California.

Parr, N.M. and Cullen, N. (1988). Risk management and Reservoir Maintenance, paper given at the Institution of Water and Environmental Management Annual Symposium 1988 Risk Management in Water and Environmental Services, London.

Penning-Rowsell, E.C., Parker, D.J., Crease, D.J. and Mattison, C.R. (1983). Flood warning dissemination: an evaluation of some current practices in the Severn-Trent Water Authority area, Geography and Planning Paper No. 7, Middlesex Polytechnic, Enfield.

Province of British Columbia (1984). Guide for preparing operation and maintenance manuals for major dams in British Columbia, Water Management Branch, Ministry of Environment and Parks, Victoria.

Prugh, R.W. (1985). Mitigation of Vapor Clouds, Plant Operations Progress 4 (2), 95–103.

Sellafield Local Liaison Committee (1982). Notice to local households concerning BNFL Sellafield Emergency Arrangements, The Committee, Sellafield.

Stanbridge, E.C. (1980). Emergency Evacuation: Emergency Planning Officer's Report, London Borough of Barking and Dagenham.

Thoits, P.A. (1982). Conceptual, Methodological and Theoretical Problems in Studying Social Support as a Buffer against Stress, Journal of Health and Social Support 23, 145–159.

Tyrer, M.J. (1988). Comment on paper by Parr and Cullen, in Proceedings of Annual Symposium – Risk Management in Water and Environmental Services, London: Institution of Water and Environmental Management.

US Nuclear Regulatory Commission/Federal Emergency Management Agency (1980). Criteria for Preparation and Evaluation of Radiological Emergency Response Plans and Preparedness in Support of Nuclear Power Plants, NUREG-0654/FEMA-REP-1, REVISION 1, NUREG/FEMA, Washington.

Wheaton, B. (1985). Models for the stress-buffering functions of coping resources, Journal of Health and Social behavior 26, 352–364.

Wolensky, R.P. (1979). Toward a broader conceptualization of volunteerism in disaster, Journal of Voluntary Action Research 8, 3–4.