Application of time series models for forecasting the global temperature anomalies

Spectral analysis of the time series for average annual values globally averaged surface temperature anomaly shows presence harmonics lunar nodal cycle with a period 18.6 years,whichcan be used to predict theseries. Three models theseries were considered: autoregression AR(p), combined model – integrated moving ARIMA(p,d,q) and artificial neural network. It is shown that ARIMA(4,1,4) gives best results predicting global three years.