ESTIMATION OF ORGANIC CARBON CONTENT IN RUSSIAN SOILS USING ENSEMBLE MACHINE LEARNING; [ОЦЕНКА СОДЕРЖАНИЯ ОРГАНИЧЕСКОГО УГЛЕРОДА В ПОЧВАХ ШССИИ С ПОМОЩЬЮ АНСАМБЛЕВОГО МАШИННОГО ОБУЧЕНИ]

The study presents a modern quantitative assessment of the content of organic carbon in Russian soils, taking into account their huge variety, and reflects the understanding of individual factors regulating and controlling the content of organic carbon in soils of the country. The paper gives the results of three-dimensional modeling of organic carbon content in soils at several standard depths (0-5, 5-15, 15-30, 30-60, 60-100 cm) for the territory of the Russian Federation with 500 m spatial resolution using the ensemble machine learning. Automated predictive mapping was based on 4 961 soil horizons from 863 soil profiles, as well as on the extensive set of spatial information, including bioclimatic variables, digital elevation model and its derivatives, and the long-term averaged time series of MODIS data. An ensemble machine learning algorithm (stacking, stacked generalization and stacked regression) was used to build models of spatial and vertical distribution. The accuracy of obtained cartographic models was assessed using spatial cross-validation. The results of spatial cross-validation show lower accuracy: the coefficient of determination is 0,46, CCC - 0.63, logRMSE -0,88 (RMSE -1,41 g/kg) compared to randomize (R2ct - 0,68, CCC - 0,81, logRMSE - 0,68 (RMSE - 0,97 g/kg)). The proposed quantitative assessment is fully automated and makes it possible to reproduce the modeling and refine the results as new soil data are obtained. © 2022 Moskovskij Universitet. All rights reserved.