地质科技通报 (Jan 2024)
Gaussian mixture model in geochemical anomaly delineation of stream sediments: A case study of Xupu, Hunan Province
Abstract
Objective The correct processing and interpretation of geochemical exploration data are critical for regional mineral exploration. High backgrounds may be misjudged as anomalies or low and weak geochemical anomalies may be ignored, if a unified anomaly threshold is adopted for geochemical exploration data in lithologically complex regions due to different elemental abundances in different lithologies. Therefore, it is essential to identify geochemical backgrounds and anomalies in lithologically complex regions based on lithologic classification. Methods Here, we propose a method for delineating geochemical anomalies based on a Gaussian mixture model of factor scores. The geochemical exploration data are subjected to factor analysis after a log-ratio transformation, and then the lithologic classification is completed by the Gaussian mixture model with factor scores. Subsequently, the standardization is performed to eliminate the lithologic background, and geochemical exploration anomalies are delineated with the processed data. This method is used to the geochemical exploration data of 1:200 000 stream sediments in Xupu, Hunan Province. Results The results show that the contents of the metallogenic elements in various lithologies of the study area are partly different, and consequently, it would be unreasonable to adopt a uniform anomaly threshold. In contrast, the method advanced in this paper can accurately classify lithology, eliminate the background of different lithologies, and enhance low and weak anomalies, with the location of the anomalies corresponding to known deposits. Conclusion Hence, the Gaussian mixture model enables effective delineation of geochemical exploration anomalies in lithologically complex regions and provides certain information for further mineral prospecting in this region.
Keywords
