The Evaluation of Uncertainties in Agricultural Risk Assessments. Cheuk S. Liu, William E. Kastenberg and Donald G. Browne, Mechanical, Aerospace, and Nuclear Engineering Department, 48-121 Engineering IV, University of California, Los Angeles, CA 90024-1597
Risk assessment and risk management have become increasingly popular techniques for assisting decision-makers in their evaluation of the risks and benefits of agricultural chemicals. The analysis of environmental transport, transformation, and fate of a chemical is usually followed, under varying scenarios, from its application in the field to its human uptake, either directly from environmental media (air, water, soil, etc.) or indirectly through the food chain. Many computer models have been developed to predict a chemical's movement and environmental distribution, as well as the consequences of human exposures. However, many of the quantitative values used in risk assessments are highly variable or poorly known, often varying over several orders of magnitude. Uncertainties that must be evaluated when conducting risk assessment studies of agricultural chemicals include: 1) variable or poorly known quantities of the chemicals themselves as well as environmental variations; 2) receptor exposure and uptake variations; and 3) biological and toxicological variations. This paper examines the nature and magnitude of the 3 categories of uncertainties mentioned above in agricultural risk assessments and illustrates the effect of uncertainty propagation through the risk assessment process. The Monte Carlo simulation methodology is used to quantify the uncertainties associated with selected parameters for an example case using atrazine.