Risk Assessment, Probabilistic

Abstract A primary criticism of environmental risk assessment as it has been practiced in the last 20 years has been the failure of analysts to properly account for issues of variability and uncertainty. Many risk assessments focused on producing ‘point estimates’ of risks have provided risk managers and decision makers with information that is at best incomplete and at worst dangerously misleading. As the field of risk assessment has matured, techniques for better treatment of variability and uncertainty have been developed and incorporated into what is known as ‘probabilistic’ risk assessments. In contrast to the point estimate approach, where single values are assumed for each model input, the probabilistic approach quantifies the possible range of values that each input may take, as well as the probability associated with each value. Modern computational power then allows the analyst to generate a multidimensional distribution of risk that reflects both the variability of risk across different individuals in the population as well as the uncertainty associated with each estimate.