Predicting the leachability of pesticides from soils using near-infrared reflectance

The purpose of this study was to investigate if near-infrared reflectance (near-IR) spectroscopy and chemometric methods can be used as a rapid and simple procedure to predict the leachability of pesticides from soil samples. The intention was to replace the often tedious determination of a number of chemical and physical parameters used to characterize soils. For selection of calibration samples, near-IR spectra were obtained for 76 different agricultural soil samples. Principal component-analysis of the spectra was used to select a subset of 20 samples which were fortified with six pesticides. Extraction was performed with supercritical fluid extraction (SFE) or solid phase microextraction (SPME). Partial least-squares regression of the near-IR data could predict 76% and 77% of the matrix-analyte interactions found by the SFE and SPME methods, respectively. The results clearly indicate that near-IR can be used to characterize soils for the prediction of the leachability of pesticides.