Selection of an elemental association related to mineralization using spatial analysis

Geochemical exploration plays a significant role in mineral exploration because it can help lead to the discovery of various types of mineral deposits. For a given geochemical exploration dataset, a number of elements are typically measured synchronously. Of these elements, several are associated with mineralization that can guide exploration, and the remaining elements may reflect other geological processes or events. How to identify elemental associations that are spatially correlated with locations of mineralization using spatial statistics, was addressed in this study. To do so, a hybrid method that combines the multifractal spectrum and the receiver operating characteristic (ROC) curve was applied. The spatial distribution characteristics of geochemical patterns were first investigated through multifractal analysis, in which enriched elements can be identified with a symmetry index R of > 1. These enriched elements were further examined using ROC to evaluate the spatial correlations between geochemical patterns and locations of mineral deposits. The elements with areas under curve (AUC) and ZAUC values obtained from the ROC curve > 0.5 and 1.96, respectively, which represent strong positive spatial relationships with mineralization, comprise the identified elemental association for locating mineralization. This hybrid method provides objective selection criteria for elemental associations related to mineralization based on spatial statistics. A case study was conducted in southwestern Fujian Province, China to demonstrate the capability of the proposed method. An integrated geochemical map based on the identified elemental association related to skarn-type Fe mineralization was produced and can be employed as a reference in further mineral exploration in the study area. However, the geological environment for the formation of mineralization should be considered in conjunction with the proposed approach.