Methodological features of long-term forecasting of grain crop yield

Abstract

The article presents the results of analysis, modeling and formation of long-term forecasts of grain crop yields based on observations of the dynamics of time series of yields in the Russia and USA for a period of more than a century and a half. To increase the reliability and reliability of the forecasting results, it is proposed to use a generalized logistic function, which allows reducing the "error corridor" for large (ten or more years) forecast horizon. This is especially important in conditions of low predictability of climatic environmental factors, environmental crises, sanctions and the accompanying high volatility of yields over significant observation periods.