Rapid prediction of total petroleum hydrocarbons concentration in contaminated soil using vis-NIR spectroscopy and regression techniques

Abstract

Visible and near infrared spectrometry (vis-NIRS) coupled with data mining techniques can offer fast and cost-effective quantitative measurement of total petroleum hydrocarbons (TPH) in contaminated soils. Literature showed however significant differences in the performance on the vis-NIRS between linear and non-linear calibration methods. This study compared the performance of linear partial least squares regression (PLSR) with a nonlinear random forest (RF) regression for the calibration of vis-NIRS when analysing TPH in soils. 88 soil samples (3 uncontaminated and 85 contaminated) collected from three sites located in the Niger Delta were scanned using an analytical spectral device (ASD) spectrophotometer (350-2500nm) in diffuse reflectance mode. Sequential ultrasonic solvent extraction-gas chromatography (SUSE-GC) was used as reference quantification method for TPH which equal to the sum of aliphatic and aromatic fractions ranging between C and C. Prior to model development, spectra were subjected to pre-processing including noise cut, maximum normalization, first derivative and smoothing. Then 65 samples were selected as calibration set and the remaining 20 samples as validation set. Both vis-NIR spectrometry and gas chromatography profiles of the 85 soil samples were subjected to RF and PLSR with leave-one-out cross-validation (LOOCV) for the calibration models. Results showed that RF calibration model with a coefficient of determination (R) of 0.85, a root means square error of prediction (RMSEP) 68.43mgkg, and a residual prediction deviation (RPD) of 2.61 outperformed PLSR (R=0.63, RMSEP=107.54mgkg and RDP=2.55) in cross-validation. These results indicate that RF modelling approach is accounting for the nonlinearity of the soil spectral responses hence, providing significantly higher prediction accuracy compared to the linear PLSR. It is recommended to adopt the vis-NIRS coupled with RF modelling approach as a portable and cost effective method for the rapid quantification of TPH in soils.

Rapid detection of alkanes and polycyclic aromatic hydrocarbons in oil‐contaminated soil with visible near‐infrared spectroscopy

“…() (1712 and 1752 nm) and Douglas et al . () (1712 and 1758 nm) for hydrocarbon‐contaminated soils. The wavelength of 2207 nm is near to those reported by Chakraborty et al .…”

“…Results also showed that the PLSR model performed better for PAHs than for alkanes, with good performance in cross‐validation ( R 2 = 0.76, RMSEP = 0.81 mg kg −1 and RPD = 2.07) and fair performance in prediction ( R 2 = 0.56, RMSEP =1.21 mg kg −1 and RPD = 1.55) (Table and Figure ). The better performance of RF compared with PLSR can be attributed to the fact that the RF modelling technique typically yields better results when the relation between reflectance and concentration is nonlinear (typical in soils) (Nawar et al ., ; Douglas et al ., ), whereas the PLSR model fits only linear relations (Nawar et al ., ). Results obtained with PLSR are not as good as those already reported in the literature.…”

“…Soil sampling locations for the three contaminated sites in the Niger Delta (Bayelsa and Rivers State), Nigeria (after Douglas et al , ). [Colour figure can be viewed at wileyonlinelibrary.com].…”

Prioritisation of Crude Oil Contaminated Sites to Inform Risk Decision Making Using Soil Quality Index

“…The soil samples were collected from three oil spill sites (Site 1:6.4527 o E, 5.1542 o N; Site 2: 6.4511 o E, 5.1357 o N; Site 3: 6.1224 o E, 4.9314 o N) in the Niger Delta region of Nigeria (Figure 1). For more details on the studied area and samples used (e.g., sampling location, method for sample collection, sampling depth, mass of samples, and sample preservation) readers are kindly referred to the study [22].…”

Soil contamination sampling intensity: determining accuracy and confidence using a Monte Carlo simulation

“…Reflectance spectroscopy has been successfully used to quantitatively measure petroleum hydrocarbons in a range of soil types (Forrester et al 2013;Okparanma and Mouazen 2013;Okparanma et al 2014;Douglas et al 2018). In addition, XRF has been used to successfully measure heavy metal concentrations in soils affected by human activities (Carr et al 2008;Radu et al 2013).…”