Forecast Modeling of Monthly Runoff with Adaptive Neural Fuzzy Inference System and Wavelet Analysis

There is no real periodicity in the changes of hydrological systems. Changes in a hydrological system take place with different periodic variations from time to time. In this paper, a new method was utilized to predict monthly runoff with a wavelet analysis technique. Taking advantage of localized characteristics of wavelet transform and the approximation function of an adaptive neural fuzzy inference system (ANFIS), the combined approach of wavelet transform and ANFIS was used to predict monthly runoff. The ANFIS forecast model for monthly runoff was established based on wavelet analysis. To solve the problems related to the large amplitudes of intra- and interannual variation of monthly runoff, a resolving and reconstructing technique of wavelets was utilized to decompose runoff series into different information subspaces, by which decomposition signals with different frequencies were obtained. Based on the evaluation of simulated and measured values in Yichang Station, it was found that the percent of the pass of relative error was 100% and the effect of prediction was acceptable. The certainty factor, , was 0.91 and the prediction level was A. By comparing the measured and predicted values, it was found that with this model, the trend of originals could be forecasted, but improvements are still needed. Because the new model is sensitive to sudden changes in rainfall, combined with the irregular monthly runoff variation, it is difficult to forecast runoff with this model, which should be improved in the future. (C) 2013 American Society of Civil Engineers.