Abstract
As new pesticides are developed, regulatory agencies have a need to determine if these chemicals will degrade water quality. There is interest in determining the utility of computer models in this evaluation process. This paper evaluates the Chemical Movement in Layered Soil (CMLS) model for this purpose. The first section assesses the ability of CMLS to predict experimental results obtained in lysimeters with two pesticides and five soils. All the soil and chemical parameters needed by the model were obtained independent of the leaching lysimeters so no calibration was involved. The ability of the model to predict drainage and total chemical leached from the lysimeter are evaluated. The presence of pesticides in the leachate after very little drainage implies preferential flow was involved in many of these lysimeters. Because CMLS makes no attempt to model preferential flow, it underestimated pesticide leaching in those soils. The second section of the paper evaluates the uncertainty in model outputs due to unknown future weather at a site, due to uncertainty in pesticide properties, and due to uncertainty in soil properties. The uncertainty in amount leached is more than ten orders of magnitude. The paper illustrates how uncertainty can be incorporated into the decision‐making process. Because CMLS requires input parameters which are readily available, is easy to use, requires little computing time, and can produce probability distributions for various outputs, the model is useful for evaluating pesticide leaching in many soil‐chemical‐weather systems. However, model‐users must be aware of assumptions inherent in this or any model.