Low-cost digital mapping of soil organic carbon using optical spectrophotometer and Sentinel-2 image

Abstract

Nowadays, scientific research is involved to identifying methods for measuring and mapping some soil properties allowing the cost reduction for sampling and laboratory analyses. In the topic of precision agriculture, it is of interest to obtain accurate spatial distribution maps of soil organic carbon to drive fertilization and variable rate seeding. The aim of this work is to test the   Pro spectrophotometer using both dry and wet topsoil color for the estimation of the total organic carbon (TOC) in a geographically limited area with relatively low soil variability. The relationships obtained using multiple linear regression (R2 of 0.54, p-value < 0.001) were not very accurate due to a high variance between the measured and predicted values. However, starting from a soil geo-resistivity survey and free remote sensing images, this method has proved effective in increasing the number of measured points thus making an important contribution to the creation of an interpolated precision maps of soil organic carbon, calibrated for the study area.