Possibility of determining soil pH using visible and near-infrared (Vis-NIR) spectrophotometry

Abstract

Aim: The present study was conducted to investigate possibility of using fiber-optic visible and nearinfrared (Vis-NIR) spectrophotometry for determining soil pH. Methodology: Diffuse reflectance spectra of 272 soil samples taken at a depth of 0-20 cm from cultivation soil in Bursa Province, Turkey, were measured in the laboratory to determine soil pH. Before apartial least square (PLS) analysis of the reflectance spectra values was conducted, all of them were randomly split into calibration (70%) and validation (30%) sets. A model for the prediction of soil pH from reflectance spectra values obtained from spectrophotometry was established using PLS regression analyses with full crossvalidation. Results: The regression value (R2), root mean square error of prediction (RMSEP) and residual prediction deviation (RPD) of established model were found to be 0.69, 0.70%, and1.413,respectively. Interpretation: The results demonstrated a moderate level of success for the model in predicting soil pH.