Based on the sensitivity analysis at annual, monthly and daily scales, a multi-step optimization scheme is designed to optimize parameters in Xinanjiang model with the help of SCEM-UA algorithm. In this scheme, parameters are classified into three groups and optimized group by group at different time scale. In this paper, the attention is focused on the parameters for data adjustment, Cp and Cep, which are optimized in the first step. Instead of optimizing Cp and Cep directly in two dimensional spaces, the concept of aridity index (the ratio of annual potential evaporation to precipitation) is introduced into the optimization process, and proposed a relationship describing Cp and Cep as a function of annual runoff coefficient and pan aridity index (the ratio of annual pan evaporation to precipitation). In this way, the feasible parameter space is narrowed and efficiency of sample points which generated for optimization in SCEM-UA algorithm is improved. In addition, it provides a simplified and quick way to get acceptable values of Cp and Cep when the request of precision is not so strict.
2013 Japan Society of Civil Engineers