МОДЕЛИРОВАНИЕ МНОГОУРОВНЕВОЙ ДИНАМИКИ УРОЖАЙНОСТИ СЕЛЬСКОХОЗЯЙСТВЕННЫХ КУЛЬТУР

Abstract

Listen

Translate article

The objective of the study is to develop an algorithm for modeling crop yields based on the dynamic-stochastic and cyclical properties of long-term characteristic series. Tasks: identifying the properties of time series of agricultural crop bioproductivity based on their consideration as multi-level structures with cyclical fluctuations; using the properties of variability of long-term characteristic series to build forecasting or stochastic assessment models; implementing the modeling algorithm using the example of grain and leguminous crop yields in Russia and wheat bioproductivity in the USA. The object of the study is long-term variability of agricultural crop yields. An algorithm for constructing a dynamic-stochastic multilevel model for forecasting the variability of agricultural crop yields under different conditions of a combination of the state of the natural environment and technological processes, taking into account cyclical fluctuations of the characteristic, has been created. The algorithm has been implemented for modeling empirical series. When describing the long-term variability of the characteristic and assessing favorable and unfavorable events, trends, histograms of the distribution of relative frequencies of local minimum and maximum cycles, and probability distributions of yield losses and gains have been used. Multilevel modeling allows assessing the dynamics of agricultural crop bioproductivity under different conditions of producers' activities: determining losses in unfavorable conditions of the natural and technological environment, identifying yield gains in favorable situations. Using stochastic methods, it is proposed to estimate rare events corresponding to very high and very low crop yields, as well as transitions of series levels into events. Based on the identified multi-level trends and the frequency of local extreme cycles, a retrospective forecast of the characteristic with a lead time of 3 years is obtained.