Long-term Crop Yield Forecasting in the Urals Steppe Zone Using Modern Methods for the Estimation of Solar—Terrestrial Relations

摘要

Present-day knowledge of the rhythmic pattern of solar—terrestrial relations was applied to forecasting of the crop yield in the steppe zone of the Urals. Crop yield estimates were obtained using the neural network method, which involves multilayer perceptrons in time series forecasting, and the method of residual deviations combined with the epoch-folding method. Encouraging results were obtained after three years of experiments. Timely forecasts allow billions of rubles to be saved in the Orenburgoblast alone through the conservation of energy resources in drought years.