Abstract of Meeting Paper

Society for Risk Analysis - Europe 1998 Annual Meeting

Methodology for Probability Assessment of Large-Scale Emergencies. Dr. V. A. Akimov, EMERCOM of Russia, Davidkovskaia Str., 7, Moskow, 121352, Russia, telephone/fax (7095) 443-24-85, e-mail kuzmin@quest.net.kiae.su

Methodology for statistical treatment of data on accidents and probability prognosis of large-scale emergencies, i.e., emergencies characterized by essential impact and probability of their annual occurrence but not by their number has been presented.

The methodology making it possible to obtain the generalized statistical data on the emergency number and impact has been analyzed. An approach on comparative assessment of statistical data accumulated during a number of years on the basis of theory of similarity has been given. Relations for calculation of numeral characteristics for identification of emergencies according to their impact have been derived.

The methodological efficiency taking into consideration the death-rate and the number of the individuals affected by explosions and fires within the Russian territory has been verified.

The methodology under study ensures to develop the mathematical models and software allowing to establish the statistical regularities on the dynamics number of emergencies and their classification according to impact/ The availability of the methodology will provide a substantiated prognosis of hypothetic large-scale emergencies for the next year. This is an important condition of planning the provisions and measures to control the comprehensive risk for the population that results from natural and manmade disasters and mitigate their impact.

The methodology (by combination of well-known statistics of emergencies accumulated during several years) ensures to improve substantially the prognostic accuracy. In the report it has been demonstrated that proceeding from the statistics of those who died and suffered as a result of explosions and fires in the Russian territory for 1991 - 1996 period, the prognostic accuracy has been improved twice.

It has been concluded that for effective risk management in emergency and mitigation of impact a short-term (operative) and long-term prognoses of the disaster number and probability as well as the disaster impact are required. To solve this problem a support system for decision-making is to be developed. The latter system is to be based on mathematical models for analysis of statistical data on disasters and prognosis as well as data bank for diverse kinds of emergencies, the causes and regions of their occurrence being taken into consideration. Apart from timely and accurate prognoses, the latter system will provide to identify the disaster according to its structure and range the regions and causes in compliance with their potential hazard for subsequent effective control and management of the population safety.