Predicting Soil Organic Carbon and Total Nitrogen in the Russian Chernozem from Depth and Wireless Color Sensor Measurements

Abstract

Color sensor technologies offer opportunities for affordable and rapid assessment of soil organic carbon (SOC) and total nitrogen (TN) in the field, but the applicability of these technologies may vary by soil type. The objective of this study was to use an inexpensive color sensor to develop SOC and TN prediction models for the Russian Chernozem (Haplic Chernozem) in the Kursk region of Russia. Twenty-one dried soil samples were analyzed using a Nix Pro™ color sensor that is controlled through a mobile application and Bluetooth to collect CIEL*a*b* (darkness to lightness, green to red, and blue to yellow) color data. Eleven samples were randomly selected to be used to construct prediction models and the remaining ten samples were set aside for cross validation. The root mean squared error (RMSE) was calculated to determine each model’s prediction error. The data from the eleven soil samples were used to develop the natural log of SOC (lnSOC) and TN (lnTN) prediction models using depth, L*, a*, and b* for each sample as predictor variables in regression analyses. Resulting residual plots, root mean square errors (RMSE), mean squared prediction error (MSPE) and coefficients of determination (R2, adjusted R2) were used to assess model fit for each of the SOC and total N prediction models. Final models were fit using all soil samples, which included depth and color variables, for lnSOC (R2 = 0.987, Adj. R2 = 0.981, RMSE = 0.003, p-value < 0.001, MSPE = 0.182) and lnTN (R2 = 0.980 Adj. R2 = 0.972, RMSE = 0.004, p-value < 0.001, MSPE = 0.001). Additionally, final models were fit for all soil samples, which included only color variables, for lnSOC (R2 = 0.959 Adj. R2 = 0.949, RMSE = 0.007, p-value < 0.001, MSPE = 0.536) and lnTN (R2 = 0.912 Adj. R2 = 0.890, RMSE = 0.015, p-value < 0.001, MSPE = 0.001). The results suggest that soil color may be used for rapid assessment of SOC and TN in these agriculturally important soils.