Abstract:
Spectral analysis and detection of abrupt changesinhydrological variables are receiving considerable attention to understand the impacts of climate change and socio-economicchanges.Spectralanalysisprovides better understanding of the change in pattern of the prominent climatic variables in a river basin. Precipitation is a prominent factor of the hydrologic cycle and has a direct influence on the water resources of a region. If the precipitation pattern changes,it affects soil moisture, groundwater reserve and streamflow. Thus, proper understanding of spectral analysis and correlation of precipitation and its change in pattern will provide better management and planning of water resources in a region. In this study, the surrounding area ofSrikakulamdistrict of Andhra Pradesh which is prone to flood due to River Nagavalihas been studied.Daily datasets of rainfall (1990 to 2009) and discharge (1990 to 2009) are analysed to understand their time frequency behaviour. In order to help hydrological aspect, the analysis of rainfall and discharge data was doneusingauto-correlation, cross-correlation, fast Fourier transform and wavelet transform. The fast Fourier transform and wavelet transform were applied to these time series in order to determine dominant spectral components.The cross-correlation coefficient values are showingmaximum after the time delay of 1 day and 2 days, which relates to the time delay of 1 day to 2 days between the rainfall and discharge time series. The autocorrelation and fast Fourier transform (FFT) power spectrum show that the periodicity for both the discharge and rainfall time series data is 372.4 days.The wavelet spectrum shows the periodicity of 370 days approximately for both the rainfall and discharge time series.
