Cu Mineralization Anomalies Information Extraction And Non-stationary Relationships Researching Of Metallogenic Factors In Gobi Desert Covered Area Of The Eastern Tianshan Region,China

Eastern Tianshan Mountain region is located in the combinative area of Kazakhstan-Junggar Plate and surrounded by Junggar Plate and Tarim Plate.As an important metallogenic prospect area,Eastern Tianshan Mountain region is riched in many resources,such as Oil,Coal,Uranium,Iron,Cu,Ni and Au.However,due to the special natural and geographical conditions,the surface area are covered by the gobi and desert and cause the information related to mineralization was weakened seriously.Some scholars represented by Qiuming Cheng introduce the nonlinear theory into mineral potential mapping and directly and efficiently solved the issues mentioned above.In this paper,we carried out geochemical and mineral anomaly extraction by the method of Singularity Index,Spatial-Weighted Principle Component Analysis and Geographically Weighted Regression based on systematic collection of geological,geochemical and remote sensing materials of study area and evaluate the spatial non-stationary of Cu mineralization and its controlling factor aim to provides the scientific evidence for the mineral potential mapping in this area.Firstly,the basic principle and process of the mathematical models in this paper were briefly summarized.We suggest that the Cu mineralization in the study area was controlling by four factors: the weak anomalies of Cu,the faults system,the mafic rocks and the alteration,based on the research of the characteristic and genetic mechanism of the Cu mineralization.A series of geochemical element combinations were screened according to their relativity with the controlling factors and were calculated by Principle Component Analysis regression.The newly calculated principle component with large eigenvalues was selected as the representative of the controlling factors of mineralization.The visible–infrared reflection of Aster data was used to extract the altered minerals anomalies information,while the thermal emissivity data were used to extract the information of mafic rocks.Combining with the geochemical analysis,the result enhances the identification of the four controlling factors and provided strong evidence for the evaluation of the mineralization prospecting areaIn this paper,both the global and the local regression model were used to describe the mathematical relationship of the Cu mineralization and its controlling factors.The result suggests that the Geographically weighted regression model is better than the global model in accuracy and stability.What's more,the coefficients of the variables,offered by Geographically weighted regression model,were calculated popint-by-point,which is mean that the contribution of the variables is varied by the change of spatial location.This just perfectly explains the spatial non-stationary relationship between the Cu mineralization and its controlling factors.Finally,we defined a weighted factor according to the metallogenic regularity of Cu-Ni sulfide deposit in the study area and estimate five prospecting target area of Cu mineralization based on the analysis of Principal Component Analysis and Spatially Weighted Principal Component Analysis.