A new forecasting model for nonstationary environmental data Original Research Article

The object of the present study is to develop a new forecasting model for the atmospheric temperature of the continental United States. We shall analyze the pattern of the temperature time series, and illustrate the usefulness of the duplicated mean of the signal. Removing the duplicated mean time series from the original temperature recording series simplifies the forecasting process. The accuracy of this proposed methodology will be demonstrated in comparison with the classical multiplicative Autoregressive Integrated Moving Average, ARIMA, model that is often used.