Potato yield forecasting based on a fuzzy-possibility model

Abstract

Potato production is a complex dynamic system that requires ensuring the stability of its functioning in changing, uncertain economic conditions and a scientifically based approach to making effective management decisions. The purpose of the research was to assess and predict potato productivity depending on the level of mineral nutrition and the length of the growing season using expert knowledge and a fuzzy-possibility approach in the conditions of the Republic of Karelia. The experiments were carried out in 2019-2021 on three potato varieties of different early maturity: Udacha, Red Scarlet and Aurora on soddy-weakly podzolic sandy loamy well-cultivated soil. The microfield experiment included two options for applying mineral fertilizers: N80P60K120 и N120P90K180, the variant without fertilizers served as a control. Studies have shown that the application of mineral fertilizers significantly increases the yield of tubers by 10-58%. The yield of potatoes without the use of fertilizers averaged 28.4 t/ha in the experiments. The highest potato yield was obtained on the early variety Udacha (38.4 t/ha) against the background of the use of mineral fertilizers at a dose of N80P60K120. An adequate fuzzy-possibility model (FPM) based on explicit and implicit expert knowledge has been synthesized to predict potato yield. It is shown that the most significant factors in assessing the yield of potatoes of all tested varieties are the level of mineral nutrition (X2) and soil fertility (X5). The synthesized FPM can be used for the planned potato yield under certain external and similar technical and technological conditions.