Visible and Near-Infrared Spectroscopy Analysis of a Polycyclic Aromatic Hydrocarbon in Soils

Table 2

Summary of calibration results for phenanthrene obtained by partial least-squares (PLS) cross-validation analysis carried out with spectra and chemical variables of 25 samples for various concentrations of diesel and moisture and clay contents.

Diesel conc. (mg kg^−1)	Reflectance spectra					Category^a	First derivative spectra					Category^a
	𝑟^2	RMSE (mg kg^−1)	SD	RPD	LV		𝑟^2	RMSE (mg kg^−1)	SD	RPD	LV
30,000	0.86	0.11	0.30	2.77	2	A	0.84	0.12	0.30	2.53	2	A
60,000	0.75	0.18	0.37	2.06	2	A	0.74	0.19	0.37	1.97	2	B
90,000	0.89	0.17	0.54	3.11	4	A	0.93	0.14	0.54	3.88	4	A
120,000	0.50	0.36	0.51	1.42	2	B	0.46	0.38	0.51	1.36	2	C
150,000	0.81	0.20	0.46	2.33	2	A	0.77	0.22	0.46	2.11	2	A
Field-moist^b	0.90	0.16	0.52	3.18	6	A	0.86	0.18	0.52	2.85	6	A

aCategory of prediction (cross-validation) is the ability of PLS regression analysis for parameter validation and prediction.
Criteria: excellent (A) if RPD > 2.0, almost good (B) if 1.4 ≤ RPD < 2.0, and unreliable (C) if RPD < 1.4 [27].
^bField-moist (moisture content = 9.04–16.13%; clay content = 9–74%; diesel concentration = 30,000–150,000 mg kg⁻¹).
LV: latent variable; RMSE: root-mean-square error; RPD: residual prediction deviation; SD: standard deviation.