In modeling periodic hydrologic series, the amplitudes and phase angles of the harmonics reflect the magnitudes and timing of the seasonal peaks. Demodulation‐remodulation captures the temporal variations in these stochastic parameters, thereby providing useful information and more accurate models than traditional Fourier series models. Unfortunately, this technique has been neglected by hydrologic modelers. In this paper, application of the technique to a synthetic series validates the theory. The technique is then applied to the streamflow record of the Uda Walawe River which exhibits a strong bimodal annual pattern. Fourteen models are developed, providing a comparison between moving average filters commonly used in demodulation studies, the Fourier filter used in Fourier series analysis, and composite models which incorporate both filters. In addition, the development of a prototype forecasting filter demonstrates the viability of using demodulation‐remodulation for forecasting.
ASJC Scopus subject areas
- Water Science and Technology