Use of geoinformation and neurotechnology to assess and to forecast the humus content variations in the steppe soils

Abstract

This paper reports the results obtained using the systemic basin approach, geoinformation, and neurotechnology for modeling and forecasting of the humus spatial inhomogeneity and content variations in the steppe and dry steppe zones (Kherson oblast, Ukraine). The general trend of such variations has been determined in the 0–40 cm layer for 42 years. The intensive use of irrigation and drainage activities in 1970–1989 resulted in a significant humus depletion by 0.36% on average (from 2.56% to 2.2%). The analysis in 4450 observation points has yielded a decrease in the variability, the rising polynomial dependence of the humus enrichment from the west to the east, and the logarithmic dependence from the south to the north. The neurotechnological modeling has made it possible to develop the artificial neural network for the spatiotemporal modeling of the humus content in the soils. The humus is predicted to be subject to the irreversible process of gradual depletion in the 0–40 layer until 2025 upon the use of the existing agrotechnologies: rainfed land by 0.01%/year and irrigated land by 0.03%/year. This result defines the territorial priorities of the regional policy and suggests the differentiated efficiency evaluation of the soil-protective unit of the farming systems.