Analyses of Rainfall Time Series in the Loess Plateau of China - Periodical Fluctuation and Links with Sea Surface Temperature -

抄録

　Periodical fluctuation and links with sea surface temperature (SST) of rainfall time series in the Loess Plateau of China are discussed in this study. The frequencies of the rainfall time series were obtained by spectral analysis and applied for Fourier series to fit the original time series. The Akaike Information Criterion (AIC) was used to optimize combinations of the frequencies. Correlations of the inter-annual rainfall time series in July and August with SST over the South China Sea in the previous winter indicated significant links. Forecasting of the rainfall in July and August was carried out by the artificial neural network (NN) from SST over the South China Sea.