Application of classical statistics, logratio transformation and multifractal approaches to delineate geochemical anomalies in the Zarshuran gold district, NW Iran

作者：Nazarpour Ahad^*; Omran Nematolah Rashidnejad; Paydar Ghodratolal Rostami; Sadeghi Behnam; Matroud Fatemeh; Nejad Ali Mehrabi

来源：Geochemistry, 2015, 75(1): 117-132.

DOI：10.1016/j.chemer.2014.11.002

• Geochemical anomaly
• Logratio transformation
• Fractal
• Zarshuran

The aim of this study is to discriminate the geochemical anomalies in the Zarshuran district, NW Iran, using different geochemical methods and present a more useful method where anomalous areas better coincide with the geological features. For this methods of delineation, geochemical anomalies were compared using geological features, occupied area of anomalies respect to the total study area, and field observations. Frequency based analysis such as mean + 2SDEV and median + 2MAD and concentration-area (C-A) multifractal methods were adopted for estimating thresholds and separating geochemical anomalies in uni-element data, as well as multi-element ones. Threshold values obtained from mean + 2SDEV and median + 2MAD, from original point geochemical data, are smaller than those of the pixel values; this may be due to the stronger variance of pixel values. In addition, the C-A multifractal method, as a useful tool to identify weak geochemical anomalies, was applied for defining the threshold values. Robust principal component analysis (RPCA) methods coupled with isometric log-ratio OW transformations were utilized to open the geochemical data in order to reduce the effects of the data closure problem. The 20-quantile intervals decomposed anomaly maps from PC1 were obtained from the classical PCA, robust PCA showed that the upper quintile (>80 quintile) of classical PCA covers a larger area (32.54%) than the robust PCA (18.16%), and as a result, the robust PCA displayed smaller areas and has good spatial associations with outcrops of hydrothermal Au-As mineralization in this area; coincident with the known Zarshuran former mining area (ore field), Zarshuran unit, Ghaldagh silicified limestone occurrence and newly explored works confirmed by field observation. Although the C-A model shows a smaller area (8.06%), this anomaly location is limited to the Zarshuran old mining area with no new exploration targets. Comparison of the models indicates that the RPCA model is not only beneficial to further Au exploration in the study area, but also provides a meaningful geological study to the community of the compositional data analysis.

• 出版日期2015