Assessment of Geochemical Anomaly Uncertainty Through Geostatistical Simulation and Singularity Analysis

摘要

Geochemical anomalies are commonly separated into different geochemical anomaly levels based on one or more thresholds. However, this practice may cause some important geochemical anomaly information to be lost and subsequently draw wrong decisions for mineral exploration. In addition, previous studies indicate that sparse geochemical sampling always entails uncertainty resulting from conventional geochemical interpolation methods because of smoothing effect. Uncertainty can propagate through the various steps of geochemical data analysis that may lead to significant impact on the final results (e.g., anomaly interpretation and mineral exploration). For geochemical anomaly identification, quantifying the probability of unsampled locations and characterizing the spatial uncertainty of geochemical anomaly based on (not) exceeding a key threshold is very important for practical demands such as exploration risk assessment. Considering the limitations of deterministic modeling method and geochemical anomaly assessment, this study proposes a new method of geochemical anomaly uncertainty assessment by combining geostatistical simulation and singularity analysis. A case study for Au anomaly uncertainty assessment is presented in the west Tianshan region (China) so as to verify the feasibility and effectiveness of the proposed method. The sequential Gaussian simulation was adopted to generate a set of equiprobable realizations that were subsequently employed to produce a series of corresponding singularity index realizations by means of singularity analysis. Critical thresholds of E-type singularity index () were determined by the method of singularity-quantile plot analysis, which were used to simulate the spatial uncertainty of Au anomaly in the study area. The results show that the risk probability of Au anomaly characterized by (not) exceedance of a critical threshold can be considered as an important reference for exploration decision-making and risk management.