Conference Dates

June 18-21, 2006


Warning systems are an essential component of risk management for natural threats. They trigger active and/or passive countermeasures. This paper introduces the need for and use of warning systems through the review of recent major natural disasters (the Sumatra Tsunami, the 2005 Alpine Flood, the 2005 Hurricanes, Katrina and Rita) but also the potential flu pandemic. Positive and negative lessons are drawn from these cases. This is followed by a review of two specific warning systems which together with the preceding cases leads to a list of requirements for warning systems.

These reviews also show that it is desirable to have tools with which different warning systems and their role in risk management can be evaluated. A procedure based on decision making under uncertainty allows one to do so. The basis of this procedure and its application to warning systems, including some sensitivity analyses to demonstrate practical consequences are then shown. Decision trees and Bayesian trees are used in this context.

The paper leads to the conclusions that the basic elements of warning systems are associated with problems, which eventually may be solved. The formal risk assessment allows one to prioritize the problems and the solutions.