Application of information content to extract wavelet-based feature of rainfall-runoff process

摘要

One of the most important steps in any water resources management is the rainfall-runoff modeling. In this paper, important temporal features of rainfall-runoff process of Delaney CreekWatershed at Florida State is extracted. The rainfall and runoff time series are decomposed into several sub-series via wavelet transform and then dominant sub-series are extracted by using the Shannon entropy theory (information content). In this way, two criteria, entropy (H) and mutual information (MI) are reviewed and employed. Finally, the efficiency of above criteria is compared with the linear correlation coefficient (CC) in the issue of feature extraction of rainfall-runoff process.