A geostatistical investigation of 3D magnetic inversion results using multi-Gaussian kriging and sequential Gaussian co-simulation

Abstract

Inversion of magnetic data has a significant role in geophysical exploration and it can be utilized in the 3D (three-dimensional) modeling of subsurface bodies (e.g. geological structures, mineral deposits, etc.). In this study, Physical Property Distribution Method (PPDM) was utilized for the inversion of magnetic data and the reliability of the resultant model was later examined by means of the geostatistical methods. For magnetic anomalies, PPDM first divides the surface of interest into a large number of cells with constant dimensions and then calculates the susceptibility value for each cell. On the other hand, geostatistical methods can calculate the ore grade value at different depths of a mineral deposit. In this study, we have used Multi-Gaussian Kriging (MGK) method for estimation of iron grade (Fe %) in Darreh-Ziarat Iron Deposit. MGK is a powerful non-linear estimation method that first provides normal distribution of grade values using the plot variogram of these new values and then makes use of one of the linear estimation approaches to prepare the ore grade distribution of each block. Also, Sequential Gaussian Co-Simulation (SGCS), which is based on the normal distribution of the variables of interest, can simulate iron grade value of each block using a secondary value. Here, SGCS method was utilized as a simulation algorithm to determine the relationship between Fe and susceptibility variables. Considering the fact that magnetic susceptibility variables exist in all simulated blocks, SGCS results are highly dependent on inversion results. So, Iron grade values at different depths of Darreh-Ziarat Iron Ore deposit were calculated using MGK and SGCS methods in order to assess the correctness of inversion results. Finally, the results revealed a high degree of validity at shallower depths and a lower depth of confidence at deeper depths.