Mathematical modeling of crop yield forecast based on field monitoring and remote sensing data

A mathematical model for forecasting crop yields based on field monitoring data and remote sensing of the Earth has been constructed. The model includes the following main values: vegetation indices NDVI, total solar radiation flux at the lower boundary of the atmosphere, efficiency of using photosynthetically active solar radiation, biomass respiration costs. After the parameterization of the mathematical model using observational data, the number of uncertain (calibration) coefficients of the model decreases from 8 to 2. These coefficients are determined by the method of successive approximations when comparing the results of calculations with observational data on the yield of a particular agricultural crop (ACC) in a given field. It is shown that the values of these coefficients strongly depend on the choice of optimal conditions for the growth of ACC. To reduce the yield forecast error, an approach based on the numerical integration of the total energy flux density by methods of the second and fourth orders of accuracy is proposed. When using numerical integration methods of a high order of accuracy, the yield forecast error decreases on average by 20% compared to the widely used model for calculating biomass growth, which has the first order of accuracy.