The influence of field survey principles on the quality of approximation of the relationship between the content of soil organic matter and multi-temporal spectral characteristics

Five field surveys of the same territory were conducted to establish functional relationships between the content of soil organic matter (SOM) and multi-temporal spectral characteristics of the bare soil surface (BSS). The surveys differ in five methods for planning soil sampling sites. Three methods used traditional approaches using regular and irregular networks, and two methods used the coefficient C map of the multi-temporal (empirical) soil line (MSL). It is shown that it is difficult to obtain a stable relationship between the values of the spectral characteristics of remote sensing data and the SOM content using traditional methods. The determination coefficients of such dependencies are below 0.60, which does not allow to recommend the obtained dependencies for the purposes of SOM mapping. Organization of field work based on a pre-built map of the coefficient C of the MSL showed the possibility of establishing functional relationships between the SOM content and multi-temporal spectral characteristics of the BSS with a determination coefficient of 0.81. The reason for the different quality of regression models is the omission of a number of soil objects during traditional organization of field work and problems in positioning the sampling. Correction of traditional field surveys based on multi-temporal spectral characteristics of the BSS allows improving the quality of research and developing new methods for SOM mapping.