Effect of errors in positioning the sampling points on the assessment of the relationships between the remote sensing data and the soil properties.

摘要

The stochastic modeling of the shift in coordinates of the sampling points was performed for a typical land plot in Bryansk opolie with the purpose to assess the influence of the accuracy of the sampling point positioning on the predicted properties of the plow horizon. It was proved that, for agro-gray soils, an occasional shift of the sampling points by 2-6 m regarding the coordinates on satellite photographs of superhigh resolution may significantly change the correlation between the soil properties and the brightness in the infrared channel. Based on the modeling data, the 95% confidence interval was estimated for the possible correlation coefficients of such properties of the arable agro-gray soil as the density, moisture, specific surface, carbon, and mobile nitrogen content. The use of the brightness in the infrared channel as a covariate for building charts may noticeably improve the map's quality; however, it may unjustifiably complicate the mapping units delineation.Notes Translated from Pochvovedenie (2013) 5, 533-539 (Ru).