Ivan Vladimirovich Sobolev
-
Digital mapping of soil salinity in the southern steppe zone of Russia based on artificial neural networks and linear regressionMoscow University Bulletin. Series 17. Soil science. 2024. 4. p.170-183read more113
-
Remote sensing data is an important source of information for monitoring and mapping vegetation cover. Machine learning methods are a modern and powerful tool for data processing. However, machine learning methods combined with remote sensing data were practically not used for soil salinity assessment and mapping in the southern steppe zone of Russia. This paper examines the possibility of applying various spectral characteristics to soil salinity mapping in solonetzic complexes in the southern steppe zone of Russia (Republic of Kalmykia) using machine learning methods. A number of predictors were considered, including reflectance coefficients in blue, green, red, infrared spectral zones, vegetation indices (NDVI, NDVIt, TVI, SAVI, MSAVI, EVI1-EVI4), salinity indices (SI1-SI6), intensity indices (Int1, Int2), brightness index (BI), and an index proposed by the authors. High-resolution images from QuickBird (2.4 m) and SuperView-1 (2 m) satellites were used. Soil salinity was assessed using two indicators: specific electrical conductivity in water suspension (EC1:5) and sodium activity (aNa1:5). Two different machine learning models were applied in the study: linear regression and neural networks. According to the results obtained, the linear regression model for EC1:5 in 0–30, 0–50, 0–100 cm layers has coefficients of determination (R2) of 0.53, 0.59, 0.79 on the training sample; the test sample managed to obtain coefficients of determination of 0.49, 0.58, 0.70, respectively. The neural network model has significantly higher coefficients of determination: R2 for EC1:5 in layers 0–30, 0–50, 0–100 cm on the training sample is equal to 0.68, 0.91, 0.97; on the test sample is 0.87, 0.86, 0.88, respectively. This indicates a greater potential of this model for cartographic modeling of soil salinity. The best predictors were the following indices: NDVIt, TVI, EVI1, Int1. The study showed the potential of using the neural network model and spectral indices obtained with SuperView-1 images for soil salinity mapping of solonetzic complexes in the south of the steppe zone of Russia.Keywords: soil salinity; spectral indices; QuickBird; SuperView-1; machine learning; remote sensing data; solonetzic complexes; Caspian lowland
-