A fuzzy-based simulation method for modelling hydrological processes under uncertainty

In this study, a fuzzy-based simulation method (FBSM) is developed for modelling hydrological processes associated with vague information through coupling fuzzy vertex analysis technique with distributed hydrological model. The FBSM can handle uncertainties existed as fuzzy sets in the hydrological modelling systems, and solutions under an associated number of alpha-cut levels can be generated by solving 2(n) deterministic models. The lower reach of the Tarim River Basin in China is selected as a study case for demonstrating applicability of the proposed method. The developed model is calibrated and validated against observed groundwater elevation for four wells during the period 2000-2001, and generally performed acceptable for model Nash-Sutcliffe coefficient (R2) and correlation coefficient (R). The R2 is approximately over 0.65 and the correlation coefficient is higher than 0.90. Based on the technique of fuzzy simulation, the uncertainties of two parameters (K-H and LC) are reflected under different alpha-cut levels. The results indicate that, under a lower degree of plausibility, the interval between the lower and upper bounds of the groundwater elevation is wider; conversely, a higher degree of plausibility would lead to a narrow interval. The main effect of K-H is more significant than the effect of LC at most well sites. The proposed method is useful for studying hydrological processes within a system containing multiple factors with uncertainties and providing support for identifying proper water resources management strategies. Copyright (C) 2010 John Wiley & Sons, Ltd.