Uncertainty related to input parameters of (137)Cs soil redistribution model for undisturbed fields.

This study presents an alternative method to empirically establish the effective diffusion coefficient and the convective velocity of (137)Cs in undisturbed soils. This approach offers the possibility to improve the parameterisation and the accuracy of the (137)Cs Diffusion and Migration Model (DMM) used to assess soil erosion magnitudes. The impact of the different input parameters of this radiometric model on the derived-soil redistribution rates has been determined for a Romanian pastureland located in the northwest extremity of the Transylvanian Plain. By fitting the convection-diffusion equation to the available experimental data, the diffusion coefficient and convection velocity of (137)Cs in soil could be determined; 72% of the (137)Cs soil content could be attributed to the (137)Cs fallout originating from Chernobyl. The medium-term net erosion rate obtained with the calculated input parameters reached -6.6 t ha(-1) yr(-1). The model highlights great sensitivity to parameter estimations and the calculated erosion rates for undisturbed landscapes can be highly impacted if the input parameters are not accurately determined from the experimental data set. Upper and lower bounds should be established based on the determined uncertainty budget for the reliable estimates of the derived redistribution rates.