Yield Estimation and Crop State Assessment of Key Crops in Russia over Large Areas with Regionally-Parametrized WOFOST Model
Plotnikov D.E., Podgornova E.N., Meshalkina Y.L., Zhou Z., Elkina E.S., Kolbudaev P.A., Burtsev M.A.
// Moscow University Soil Science Bulletin, 2026. Vol. 81. No. 2. P. 189-196.
This study investigates the potential of the WOFOST crop growth imitation model for yield forecasting and crop state assessment for several key agricultural crops over large areas of Russia over a fifteen-year period (2005-2020), using soil and climatic datasets. Special attention is paid to adapting information from the Soil Map on a scale of 1:2500000 (edited by V.M.Fridland) to ensure correct performance of the model. Model parameters for regional crop species of barley, sunflower, and maize have been obtained for several hundred districts within the Russian agricultural belt, including characteristics of phenology, photo-synthesis, respiration, and distribution of assimilants. Parameterization is performed, using loss functions based on the Brent, Nelder-Mead, and Differential Evolution algorithms to minimize phenology and yield estimation errors. The initial dataset is divided into a training set and a control set at the 80:20 ratio to enable parameterization and independent accuracy assessment. The regionally parameterized models provide a good agreement between the error histograms for the training and control sets with a near-zero systematic shift, while the mean relative absolute deviation of yield estimates ranges from 20 to 26 %, depending on the crop. Furthermore, it is shown not only that the model estimates are close to the actual yield for the studied period, but also that the model forecast is capable of determining multi-year trends in the yield dynamics of the studied crops. Finally, the obtained model crop varieties were used to assess crop state in terms of the deviation of simulated yield from the yield at an optimal season, and therefore the constraction of corresponding deviation cartogram maps was performed. The results obtained may be used for decision support system of the agricultural sector, including yield forecasting and state assessment of the studied key crops over large areas, using imitation modeling, including forecasts made under various climatic scenarios.