IOCG AND IOA DEPOSIT SIGNATURES IN LIGHTS CREEK STOCK BY MAGNETITE CHEMISTRY VIA RANDOM FOREST ALGORITHM

Iron oxide-copper-gold (IOCG) and iron oxide-apatite (IOA) deposits are important sources of iron, as well as copper, rare earth elements and others. While the genesis of IOA deposits remain unclear, they often coincide spatially and temporally with IOCG deposits. One hypothesis is that IOCG deposits are at a shallower paleodepth and transition to a deeper sulfur-poor IOA deposit, leading to the idea that the two deposits are a single ore-forming system.

Magnetite is abundant in both IOCG and IOA deposits, forming under magmatic conditions and various hydrothermal conditions (sodic-calcic, potassic alteration). As a result, magnetite geochemistry has been commonly used in binary plots to discriminate between different deposit types. However, binary plots are limited in dimension and can plot iron oxides from a single deposit into a wide range of deposit types, thus having a low general accuracy. With many iron oxide distinguishing plots available, using machine learning to differentiate between IOA and IOCG has been shown to give a higher classification accuracy based on EPMA data.

For this study, a random forest algorithm (RF) was trained in R using a dataset containing over 1,100 samples from 24 known deposits from previous studies. The RF had a classification accuracy of 0.87 and was able to correctly identify 23 of the 24 deposits correctly. Contrary to the findings of previous studies, this model indicates that not only V but also Al are key elements in distinguishing between the two deposit types, with IOAs having a lower amount of Al and higher V than IOCGs. Concentrations of Mg, Mn were moderately significant, with Si and Ti being less.

The Lights Creek Stock in Plumas County, CA, historically seen as a Cu porphyry and recently reclassified as IOCG, hosts three deposits within a granitoid host rock: Superior and Engels (magnetite-dominated) and Moonlight (hematite-dominated). The testing dataset used 36 samples from the Superior mine and classified 20 of 36 samples having an IOA signature. Of those 20 IOA signature samples, 17 are stratigraphically deeper than the 16 IOCG signature samples. This indicates that the Superior deposit is a complex system which could be a transitional area between an IOA and IOCG system. Future work will incorporate iron oxide compositions from Engels and Moonlight deposits to give a clearer understanding of the relationship between the three deposits.