Monte-Carlo simulations for soil content determinations on Atlas

Authors: Kavetskiy, Aleksandr; Yakubova, Galina; Prior, Stephen - Steve; Torbert, Henry - Allen

Submitted to: Popular Publication
Publication Type: Popular Publication
Publication Acceptance Date: 4/28/2024
Publication Date: 6/7/2024
Citation: Kavetskiy, A.G., Yakubova, G.N., Prior, S.A., Torbert III, H.A. 2024. Monte-Carlo simulations for soil content determinations on Atlas. SCINet Newsletter, https://scinet.usda.gov/news/newsletter

Interpretive Summary:

Technical Abstract: The Monte Carlo Method, or multiple probability simulation, is a mathematical technique used to estimate possible outcomes of uncertain events. The Monte Carlo Method was invented by John von Neumann and Stanislaw Ulam during work on the Manhattan Project. It is a very powerful technique for studying the propagation of ionized particles into a medium and the secondary effects under irradiation. The USDA-ARS National Soils Dynamics Laboratory (NSDL) developed an in-situ method of neutron-gamma analysis for determining soil elemental content. This method is based on the registration of gamma rays issued from the studied soil upon neutron irradiation. Since each soil element issues gamma rays of specific energy, soil elemental content can be defined from the measured neutron stimulated gamma spectra. Monte Carlo computer simulations of neutron stimulated gamma spectra of different objects (e.g., soils, composts, calibration pits, and other agricultural media) is an important aspect of this research. Conducting experiments with different measurement system designs and large samples (2 m x 2 m x 50 cm) with various soil elemental contents and textures is expensive, time consuming, labor-intensive, and often physically impractical. Thus, computer simulations are useful for incorporating different factors to define their effects on the resultant gamma spectra. Developed and maintained by Los Alamos National Laboratory, MCNP software is used for analyzing the transport of neutrons (hence NP for neutral particles) and gamma rays by the Monte Carlo method (hence MC). MCNP simulations require a very large number of neutron events (on the scale of 109-1010 or greater) to create gamma spectra with an accuracy comparable to experimental measurements. By comparison, 107 neutrons per second and a measurement time of no less than 15 min are used in field experiments to acquire the soil gamma spectra. Thus, the application of high-performance computing (HPC) is suitable for these types of simulations. SCINet’s Atlas HPC cluster at Mississippi State University is currently used for this NSDL research project. MCNP computer simulations of neutron-stimulated gamma spectra on the Atlas HPC cluster make it possible to solve the following problems regarding soil content determinations: optimizing the design of the field measurement system; determining dose levels to soil and nearby environments during measurements (see Figure 1); developing methods and procedures for: developing calibration of the measurement system for carbon and moisture content determinations; correcting for soil moisture when measuring carbon content in the field; spectra deconvolution for determining soil component contents; neutron-gamma measurement for determining the carbon-to-nitrogen ratio in compost materials; assessing potassium content and chlorine contamination in the field.