LAUSR

Traditional soil organic carbon density (SOCD) measurement is time-consuming and costly, from field sampling to laboratory analysis. However, visible to near-infrared reflectance (vis-NIR) spectroscopy can rapidly estimate SOCD and thereby permit rapid, inexpensive measurements that support environmental management. Our aim was to explore the relationship between SOCD and vis-NIR spectra, with the goal of establishing a spectral SOCD estimation method. In this study, we sampled soils throughout northern China’s agro-pastoral ecotone and performed SOCD and spectral measurements. After pre-processing the data, we transformed the spectral reflectance into seven forms: the reciprocal logarithm, first-order differential, second-order differential, logarithmic first-order differential, logarithmic second-order differential, reciprocal first-order differential (RFOD), and reciprocal second-order differential. We then explored the relationship between SOCD and these indexes. We also analyzed spectral SOCD estimation models based on stepwise multiple linear regression and partial least-squares regression. We found that the spectral reflectance decreased with increasing SOCD, and the most sensitive bands for SOCD were between 745 and 840 nm. RFOD greatly improved estimation accuracy. The performance of a stepwise multiple linear regression estimation model based on RFOD provided the best fit (R2 = 0.77) and the lowest root-mean-square error (0.75 kg C m−2), and the best ratio of percent deviation (2.11). Our results suggest a high potential for fast and reliable estimation of SOCD using spectral techniques and provides a theoretical basis and technical support for rapid monitoring of SOCD in northern China’s agro-pastoral ecotone.