Near-Infrared Spectroscopy: Assessment of Soil Organic Carbon Stock in a Colombian Oxisol

Abstract

Soil organic carbon (SOC) is a property known for its influence on the physical, chemical, and biological characteristics of soils, which are essential when assessing their quality. SOC stock (SOCS) monitoring is a key task in climate change mitigation studies. However, the resources necessary to obtain the information required by these studies tend to be high. The objective of this study was to develop a model for estimating the SOCS of a Colombian oxisol using near-infrared (NIR) diffuse reflectance spectroscopy. In a sampling scheme of 70 points distributed over 248 ha, 313 soil samples were collected in five defined depth intervals of 10 cm each, from 0 to 50 cm. SOC was determined through an elemental analyzer, and bulk density (BD) by means of sampling cylinders. A NIRFlex spectrometer was used to acquire spectral signatures in the NIR range from the processed soil samples, and, together with the data measured in the laboratory, a statistical analysis was performed using partial least squares regression (PLSR) in order to calibrate the spectral models. Based on the residual prediction deviation (RPD), the root mean square error (RMSE), and the coefficient of determination (R2) of the validation groups, a highly representative model was achieved for the estimation of SOCS (R2 = 0,93; RMSE = 2,12 tC ha-1; RPD = 3,69), which was also corroborated with geostatistical interpolation surfaces and depth splines. This research showed NIR diffuse reflectance spectroscopy to be a viable technique for SOCS estimation in the study area.