Jean-Luc Wybo

Intelligent software agents for forest fire prevention and fighting

This paper discusses the design and the development of an original information system based on intelligent agent technologies and dedicated to forest fire prevention and fighting. It describes the work to be carried out in order to achieve and formalise knowledge in the field of forest fires. It presents a detailed demonstration built around two realistic scenarios, where several actors and software were implicated.

Mastering risks of damage and risks of crisis: the role of organisational learning

Mastering risks, accidents and crisis is a complex task which is achieved by a large number of stakeholders. In order to determine appropriate responses to risk-prone situations, researchers propose a classification in two categories: risks of damage and risks of crisis. Risks of damage correspond to situations that have been studied and for which preventive and protective measures have been taken by the organisation. In other words, there is a plan in place in the organisation. Risks of crisis, on the other hand, correspond to situations where there has been little anticipation and no previous experience and as a consequence, no preventive and protective measures have been taken by the organisation. There is no plan in place or the plan is inadequate or inoperable. Either of these situations may put the organisation into chaos, although chaos and confusion are more likely with risks of crisis. Special attention is given in this paper to the origins of crisis situations and emergence of ad-hoc organisational patterns among their management. A case study is presented to illustrate how an ad-hoc organisation has emerged during the management of the crisis that followed the Erika oil spill on the French coasts in December 1999. This paper presents the advantages associated with the development of organisational learning to analyse accidents and crisis: a better sharing of knowledge about identification of weaknesses and strengths, and a deeper involvement of people in risk mastering.

FMIS: a decision support system for forest fire prevention and fighting

This paper presents a decision support system for managers in charge of process monitoring, this domain being representative of rapidly increasing emergencies. Designing such a system means combining tasks: automate data processing; give the user efficient access to all relevant data; and synthesize pertinent information. Data used for decisions come from several independent sources. The authors introduce the concept of a scenario associated with every kind of information managed by this system, which describes the tasks related with this information and their scheduling. The system architecture is built around the association of a reactive database and a real-time information manager, whose task is to follow the scenarios describing the processes to trigger for all incoming data. This architecture, based on specifications and knowledge, allows autonomous behavior of the system in order to monitor, update and display all information needed for the users decisions. It uses a declarative approach, i.e. it separates what is application dependent (data, scenarios and processes) from what is generic, and can be used for other applications. The Fire Management Information System is an application dedicated to forest fire prevention and fighting, partially funded by the European Research Programme for Environmental Protection.

Chernobyl and Fukushima nuclear accidents: what has changed in the use of atmospheric dispersion modeling?

The threat of a major accidental or deliberate event that would lead to hazardous materials emission in the atmosphere is a great cause of concern to societies. This is due to the potential large scale of casualties and damages that could result from the release of explosive, flammable or toxic gases from industrial plants or transport accidents, radioactive material from nuclear power plants (NPPs), and chemical, biological, radiological or nuclear (CBRN) terrorist attacks. In order to respond efficiently to such events, emergency services and authorities resort to appropriate planning and organizational patterns. This paper focuses on the use of atmospheric dispersion modeling (ADM) as a support tool for emergency planning and response, to assess the propagation of the hazardous cloud and thereby, take adequate counter measures. This paper intends to illustrate the noticeable evolution in the operational use of ADM tools over 25 y and especially in emergency situations. This study is based on data available in scientific publications and exemplified using the two most severe nuclear accidents: Chernobyl (1986) and Fukushima (2011). It appears that during the Chernobyl accident, ADM were used few days after the beginning of the accident mainly in a diagnosis approach trying to reconstruct what happened, whereas 25 y later, ADM was also used during the first days and weeks of the Fukushima accident to anticipate the potentially threatened areas. We argue that the recent developments in ADM tools play an increasing role in emergencies and crises management, by supporting stakeholders in anticipating, monitoring and assessing post-event damages. However, despite technological evolutions, its prognostic and diagnostic use in emergency situations still arise many issues.

Spatial decision support and information management application to wildland fire prevention The WILFRIED System

This paper presents an integrated Spatial Decision Support System (SDSS), called WILFRIED, which is designed to fit the requirements of managers in charge of wildland fire prevention and fighting. To achieve decision support, the system combines tasks: it gives the user an efficient access to relevant data, synthesizes pertinent information (meteorology, topography, vegetation) and automates repetitive tasks (compute the wind field every two hours for example). The methodology that we use associates in the same computer software a set of tools (Data Base System, Geographic Information System, simulation models, user interface) in order to help the user with his decisionmaking job for complex problem solving. We present the WILFRIED system, we describe in particular the cooperation between various actors which are managed by an Information Manager linked with a Data Base Management System. Then we present the system components and its architecture.

The Role of Simulation Exercises in the Assessment of Robustness and Resilience of Private or Public Organizations

This paper deals with the organization of simulation exercises to prepare organizations to face emergencies. The original objective of such simulations is to train people to emergency procedures and devices; we raise the question of training people to face potential crisis situations: are simulations fitted to that objective? Through the observation of a number of exercises organized by private companies and rescue services, we can answer that naive interprettation of simulation results limits their benefits to the correction of gaps between rescribed and observed actions, without addressing complex organizational behaviour. We introduce a method to organize simulations that gives access to this complexity and to the resilience and robustness apacities of the organization by giving specific roles to observers. This method uses a model of the organization seen as a combination of three levels: structures, relations and meaning.

Forest fire danger assessment methods and decision support

Abstract Forest fire danger is one of the main emergencies in many countries all over the world. To reduce the consequences, forest managers or fire officers use prevention, by reducing biomass, by locating fighting teams in exposed areas or by patrolling. To fight fire efficiently, they use suitable means: trucks, planes, helicopters, fire retardants. The keypoint of prevention and fighting is the assessment of danger, which can be done from different points of view and at different time scales: from historical data, by real time monitoring or by forecasting. A survey of operational danger rating methods shows two approaches: assessment of a global index, integrating several aspects of forest fires, or assessment of a set of indexes representing individually the key factors. This paper describes a combination of assessment methods, integrated in a decision support system. Four methods are used, related to fire occurrence, assessing the frequency of fires from historical data, inflammability of dead fuels and live fuels, traducing the behaviour after ignition and fire severity, evaluating the difficulty of fighting. The integration of these methods is done using a knowledge-based approach in association with a relational database and a GIS.

Formalisation and Use of Experience in Forest Fires Management

This paper presents a study of the learning process in emergency management with an application to forest fire fighting. The first part presents the methodology used to determine the mental image of fire fighting management and to propose a model of representation of the individual experience gained during operations. The second part introduces a method to capitalise experience of fire management, which uses a prototype of GIS application. The third part of the article presents the different types of support that can be provided by a collective memory of individual experience and a reasoning method that includes this experience to provide several levels of support to forest fire managers. This study was undertaken in co-operation with organisations in Canada and Spain.

11/9: a historical perspective

This paper presents some key elements about the origins of modern terrorism that help to understand the terrorist attacks against the twin towers in New York and the Pentagon in Washington, on September 11, 2001. Initially associated with the struggle against tyrants, terrorist action has been used to support different causes, which have in common the battle against current authority.