Monitoring mineral-associated organic matter in tropical pastures using near infrared spectroscopy

Abstract

In order to confirm that the mitigation of greenhouse gas emissions could indeed be achieved by farmers, determinations of soil organic carbon (SOC) in total and stabilized fractions are essential, proving the effectiveness of the sustainable practices adopted by the farmer. In this sense, this study proposes an analytical methodology based on near infrared spectroscopy (NIRS) and partial least squares regression (PLSR) as an alternative for the measurement of stabilized and total soil organic carbon (SOC) in agricultural production systems. A set of 122 samples of four different land uses were sampled and physically fractionated. Total organic carbon for both fractions (mineral-associated stable fraction and particulated organic matter - labile fraction) were determined by the dry combustion method. In order to eliminate the extra step regarding soil fractionation to determine the stabilized SOC content, this study investigated different strategies to build the regression model based on partial least squares regression for the determination of the stabilized SOC from the total soil fraction. Two strategies presented the same accuracy as the reference method used to determine the stabilized SOC content in stabilized fraction, with root mean square error in validation of 1.47 g/dm3. These results indicate that both strategies proposed can determine simultaneously the total and stabilized SOC from the total soil fraction, thus eliminating the fractionation phase.