Dynamical Short-Term Prediction of Rain Attenuation in W Band: A Time-Series Model With Simpler Structure and Higher Accuracy

Most existing rain attenuation prediction models are tested only at 55 GHz. Few W-band models have been tested, and these models present problems associated with high complexity and cannot predict with high accuracy in the short term. A dynamical short-term prediction method is proposed here, and it has a simpler structure and higher accuracy than previous rain attenuation models and can theoretically be implemented at any station and frequency. The method uses the relationship among time series to establish an autoregression integrated moving average (ARIMA) model, conducts a stationary test, transforms the nonstationary series into a stationary series, and estimates the parameters.