A Monte Carlo-based Workflow for Geochemical Anomaly Identification Under Uncertainty and Global Sensitivity Analysis of Model Parameters.

Uncertainty associated with the identification of geochemical anomalies linked to mineralization has been a major concern in processing geochemical survey data. In this study, a Monte Carlo-based workflow, consisting of simulation-based local singularity analysis (LSA) and distance-based generalized sensitivity analysis (DGSA), was presented to identify geochemical anomalies associated with mineralization under uncertainty and simultaneously perform sensitivity analysis of model parameters. A case study of processing 1:200,000 stream sediment geochemical samples collected from the southwestern Fujian Province of China further illustrates and validates the procedure. The obtained results suggest that (i) the procedure is useful and flexible for identifying geochemical anomalies associated with mineralization under uncertainty and studying the sensitivities of model parameters, because different types of uncertain parameters and responses can be incorporated into the procedure, (ii) the window parameters are critical for applying LSA to identify geochemical anomalies related to mineralization, and spatial uncertainty associated with the gridding procedure is also important and can not be ignored, and (iii) the averaged singularity intensity map of the study area can serve as a valid indicator of Cu mineralization, while the uncertainty map of local singularity exponents can provide crucial information for decision-making in later stages of mineral exploration. [ABSTRACT FROM AUTHOR]