The influence of climatic characteristics and values of NDVI at soybean yield (on the example of the districts of the Primorskiy region)

Abstract

Abstract. The relevance of research. Soybean is one of the key crops in world agriculture; in recent years, soybean production has been actively developing in the Russian Far East. It is necessary to predict yield to solve problems associated with soybean production, including the planning of sown areas and export operations. The purpose of this study is: to determine the factors affecting yield, to establish the relationship between these indicators and yield, and to evaluate the accuracy of the model. Research methods. We examined climatic features and remote Earth sensing indicators of Khankayskiy, Khorol’skiy, Mikhailovskiy and Oktyabr’skiy districts of the Primorskiy region since 2008 to 2018. Meteorological characteristics of territories and values of vegetation index were obtained using the Vega Science system. Integral coefficients were additionally calculated and mutually correlating indicators were excluded from the regression model. The main result of the study is a multiple regression model, where yield is considered as a dependent variable, and the independent variables are: the maximum weekly NDVI, hydrothermal coefficient, duration of the growing season, average annual humidity, and aggregated temperature of the upper soil layer. Mean absolute percentage error of the model is 11.0 % for the Khankayskiy district, 4.8 % for the Khorol’skiy district, 9.5 % for the Oktyabr’skiy district, and 8.9 % for the Mikhailovskiy district. Scientific novelty and practical relevance. A regression model, which predict soybean yield, was developed. In general, the proposed model can be used to predict soybean yield, as well as to make managerial decisions at the regional level.