Mathematical methods of soil gamma spectra analysis for carbon content determination

Abstract

Conventional methods for measuring soil carbon are often time-consuming, labor intensive, and require laboratory analysis. Neutron-gamma analysis (NGA) offers an alternative for rapid (practically in-situ) and non-destructive measurement of carbon soil. This paper describes the developing mathematical methods of the soil gamma spectra analysis for determination of carbon levels in stimulated spectra of soil samples. Monte-Carlo simulations were carried out to create soil neutron stimulated gamma spectra using MCNP6.2 code. Two categories of analysis are investigated: peak fitting, and endmember fitting, a third method is introduced that combines the two into a single result. Evaluation of their effectiveness determines that the multivariate model performed best overall. Developed method can be used for future testing in real NGA carbon soil measurements.