Comparative Analysis Study Of The Methods For Middle-Large Scale Geochemical Combination Anomaly

Exploration geochemistry is an important prospecting method technology incontemporary geology world, and geochemical data processing and analysis is one ofthe core contents in exploration geochemical field, its purpose is reasonable andeffective to extract elements anomaly from geochemical background, which is todiscover mineralization anomalies related to ore-forming from massive data, and thencombine with synthetic information of regional geology, geophysics, remote sensinggeology to guide prospecting. Analysis and visualization research of massivegeological spatial data is an important research content for digital geological science.In this paper, compare two methods of geochemical spatial quantitativecombination anomaly, namely the factor Pan-Kriging method and multiple typicaltrend surface method, accompanied by geology tectonic、magmatic intrusive body andvolcano rock series etc. geology background analysis of research area, and establishcombination anomaly models suitable for different geology background. The factorPan-Kriging method is to combine factor analysis linear model with Pan-Krigingmethod, constructing a kind of anomaly model to reflect correlation, structure andsingularity of geochemical data, its main principle is through the correlationcoefficient between the sample data variable, and varimax orthogonal rotationprincipal factor solution, to reach the effect of variable space dimensionality reduction,form a new combination variables system, namely the elements combination,applying the drift function in Pan-Kriging method to simulate factor score distributioncurve to complete structure analysis, and give its trend and locality, finally gains combination anomaly value; multiple typical trend surface method is from anotherperspective to emerge the spatial correlation and structure of multidimensionaloriginal variables, the main principle is to seek the linear combination of minimumdimension for the original variables, combining the corresponding coordinate to buildtrend surface, according to both maximum correlation conditions, strive for theelement combination anomaly.In this paper, research area of the typical case are Heilongjiang ProvinceHuzhong District Songheyidongshan area and Jilin Province Baishan City Xichuanarea, data selected fully considers connect transitional relationship of large scale andmiddle scale, namely the boundary effect of large scale modeling and regionalmodeling and division of sampling type, which is from the middle scale streamsediment sampling area and large scale secondary halo sampling area. First, for eachresearch area at the same time apply two kinds of methods for the factor Pan-Krigingand multiple typical trend surface, respectively calculate the element combination,obtain the corresponding combination anomaly information; secondly, give precisionanalysis and scale classification of combination anomaly; finally, by interactiveoperation principle based on DEM and surfer platform, the main mineralization factoranomaly information gained by two models on the surfer platform creates twodimensional and three dimensional image about main mineralization factor, overlapthe two dimensional image of geochemical combination anomaly with thecorresponding regional geological map, meanwhile, overlap the three dimensionalcombination anomaly image with two dimensional contour image, carry out thecomprehensive geological interpret and comparative analysis, make the anomalycalculation results more vivid, intuitive.Through two methods to obtain the calculation process of combination anomalyinformation and comparative analysis result, gets a conclusion, for the HeilongjiangProvince Huzhong District Songheyidongshan area, the factor Pan-Kriging methoddelineating anomaly range in great degree coincides with the known molybdenum orelocation; for Jilin Province Baishan City Xichuan area, multiple typical trend surfacemethod delineating anomaly range relatively coincides with known gold ore bodylocation.Through the study of this paper, shows that the factor Pan-Kriging method ismore suitable for anomaly study of middle scale stream sediment sampling area,reflecting model is mostly linear measurement model under the stationary distribution conditions; multiple typical trend surface method is more suitable for anomaly studyof large scale secondary halo sampling area, reflecting model is mostly the nonlinearmeasurement model under non-stationary conditions. These two methods caneffectively confirm the geochemical element combination anomaly, and hint closelyrelated regularity between the anomaly and mineralization type, provide informationfor regional prospecting.