Discrimination of Soil Samples Collected from Haryana (India) Using Non-destructive ATR-FTIR Spectroscopy Coupled with Multivariate Statistical Analysis

Abstract

Objective: To discriminate and classify soil samples collected from different regions of Haryana, India. Methods: Attenuated Total Reflectance Fourier Transform Infrared (ATR-FTIR) spectroscopy with multivariate statistical tools is employed. A total of 232 samples were collected. A composite mixture of all districts was prepared, having twenty-nine top and twenty-nine depth soil samples. Chemometric methods, namely, PCA (Principal Component Analysis) and PCA-LDA (Principal Component Analysis-Linear Discriminant Analysis) were used to interpret the data. Findings: Soil samples are well characterized by their organic and inorganic contents. Sample clustering due to similarity in chemical composition was visualized using PCA. PCA-LDA resulted in 100% classification accuracy for top soil and 98.85% classification accuracy for depth soil. Blind test validation was carried out, which resulted in 100% and 80% prediction accuracies for top soil and depth soil respectively. The present research methodology effectively discriminated soil samples and can be utilized by forensic investigators dealing with cases that involve soil as vital evidence. Novelty: Study reveals novel unexplored geographical location, local soil variability, practical implications of non-destructive analytical technique combined with chemometrics methods, contextualization with the previous studies and the potential policy field relevance. Keywords: Soil forensics, ATR­FTIR, PCA, LDA, Discrimination