Prediction of soil organic carbon with different parent materials development using visible-near infrared spectroscopy

Jinbao Liu; Jichang Han; Yang Zhang; Huanyuan Wang; Hui Kong; Lei Shi

doi:10.1016/j.saa.2018.06.018

Prediction of soil organic carbon with different parent materials development using visible-near infrared spectroscopy

Spectrochim Acta A Mol Biomol Spectrosc. 2018 Nov 5:204:33-39. doi: 10.1016/j.saa.2018.06.018. Epub 2018 Jun 5.

Authors

Jinbao Liu¹, Jichang Han², Yang Zhang¹, Huanyuan Wang¹, Hui Kong¹, Lei Shi¹

Affiliations

¹ Shaanxi Provincial Land Engineering Construction Group, Xi'an 710075, China; Key Laboratory of Degraded and Unused Land Consolidation Engineering, The Ministry of Land and Resources, Xi'an 710075, China; Institute of Land Engineering and Technology, Shaanxi Provincial Land Engineering Construction Group, Xi'an 710075, China; Shaanxi Provincial Land Consolidation Engineering Technology Research Center, Xi'an 710075, China.
² Shaanxi Provincial Land Engineering Construction Group, Xi'an 710075, China; Key Laboratory of Degraded and Unused Land Consolidation Engineering, The Ministry of Land and Resources, Xi'an 710075, China; Institute of Land Engineering and Technology, Shaanxi Provincial Land Engineering Construction Group, Xi'an 710075, China; Shaanxi Provincial Land Consolidation Engineering Technology Research Center, Xi'an 710075, China. Electronic address: 1498716642@qq.com.

PMID: 29902769
DOI: 10.1016/j.saa.2018.06.018

Abstract

The storage of soil organic carbon (SOC) should improve soil fertility. Conventional determination of SOC is expensive and tedious. Visible-near infrared reflectance spectroscopy is a practical and cost-effective approach that has been successfully used SOC concentration. Soil spectral inversion model could quickly and efficiently determine SOC content. This paper presents a study dealing with SOC estimation through the combination of soil spectroscopy and stepwise multiple linear regression (SMLR), partial least squares regression (PLSR), principal component regression (PCR). Spectral measurements for 106 soil samples were acquired using an ASD FieldSpec 4 standard-res spectroradiometer (350-2500 nm). Six types of transformations and three regression methods were applied to build for the quantification of different parent materials development soil. The results show that (1)the basaltic volcanic clastics development of SOC spectral response bands located in 500 nm, 800 nm; Trachyte spectral response of the soil quality, and the volcanic clastics development at 405 nm, 465 nm, 575 nm, 1105 nm. (2) Basaltic volcanic debris soil development, first deviation of maximum correlation coefficient is 0.8898; thick surface soil of the development of rocky volcanic debris from bottom reflectivity logarithm of first deviation of maximum correlation coefficient is 0.9029. (3) Soil organic matter content of basaltic volcanic clastics development optimal prediction model based on spectral reflectance inverse logarithms of first deviation of SMLR. Independent variable number is 7, Rv² = 0.9720, RMSEP = 2.0590, sig = 0.003. Trachyte qualitative volcanic clastics developed soil organic matter content of the optimal prediction model based on spectral reflectance inverse logarithms of first deviation of PLSR. Model number of the independent variables Pc = 5, Rc = 0.9872, Rc² = 0.9745, RMSEC = 0.4821, SEC = 0.4906, forecasts determine coefficient Rv² = 0.9702, RMSEP = 0.9563, SEP = 0.9711, Bias = 0.0637.

Keywords: Basalt; Deviation; Hyperspectral prediction model; Soil organic carbon; Trachyte.