Modeling the washoff of 90Sr and 137Cs from an experimental plot established in the vicinity of the Chernobyl reactor.

After the Chernobyl event, a large area of land was contaminated following the deposition of radionuclides. This area became a continuing source of radionuclides to natural waters and aquatic ecosystems. In 1986, an experimental plot was constructed in a contaminated area near the Chernobyl Nuclear Power Plant to study the washoff of radionuclides by surface runoff. Concentrations of 137Cs and 90Sr were measured in the top 10 cm of the soil prior to the experiments. During two separate experiments, intense artificial rainfall was applied to the plot. A washoff scenario was then prepared with site-specific information on initial soil contamination, duration and quantities of rainfall and runoff, physicochemical properties of the topsoil, and some climatological data. Modelers were asked to predict (a) the vertical distributions of the initial concentrations of 137Cs and 90Sr in various chemical forms in the topsoil, (b) concentrations of these radionuclides in various chemical forms in the runoff water during each experiment, and (c) the total amounts of these radionuclides that were washed off during each experiment. Stochastically generated local rainfall data were used in a water budget model to generate annual average runoff and infiltration rates. A vertical, one-dimensional, multiphase, multispecies transport model was then developed to simulate the movement of contaminants in the topsoil during the 160-d period between the Chernobyl event and the experiments as well as the washouts of contaminants by runoff during the experiments and during the 24-h period thereafter. The model provided very good predictions of the vertical distributions of total contaminant concentrations in the top 10 cm of the soil; however, the concentrations in individual chemical forms were not predicted as accurately. Initially, the model overpredicted the washout of contaminants for the two experiments and the 24-h period thereafter. Fraction of runoff that flows as interflow and average sediment loading in the runoff were identified as parameters responsible for the overprediction. Calibration of the interflow fraction and adjustment of the average sediment loading in runoff to a level representative of Eastern Europe considerably improved these predictions. The complete modeling approach and comparisons of model predictions with measurements and with predictions from other modelers are presented.