This paper estimates the uncertainty in the outputs of a dust-transport model due to physical parameterizations (dry deposition) and input data (source intensity and wind fields). Sensitivity experiments are generated from a reference simulation in which one parameter is changed at a time. Comparisons of the simulations and observations allow us to access the impact of each process and the robustness of the model. The case study is a 10-day simulation of dust concentrations over East Asia during 15-24 March 2002. Results show that there is a higher uncertainty due to the input data (source intensity and wind fields) than the dry deposition velocity. Moreover, for this case, there is higher uncertainty due to the source intensity and the dry deposition velocity in the middle area (95°E-110°E) than in other areas, and there is higher uncertainty due to surface wind fields than higher-level ones.