Spatial and scaling modelling for geochemical anomaly separation

A spatial and scaling approach with a user-friendly windows program is intr oduced which can be used to assist exploration geologists and geochemists i n geochemical data analysis and anomaly separation. It can also be used for image enhancement and classification. Statistics are calculated and optimi zed within a variable-sized moving window centred at an arbitrary sample lo cation. The moving window has both variable size and shape determined by th ree parameters: r (size), beta (ratio of long and short axes), and theta (o rientation). It calculates five optimal indexes for each sample location: t he optimal statistic U(r(0), B-0, theta(0)), optimal size r(0), shape index es beta(0) and theta(0), and sealing index alpha (singularity exponent). Th ese indexes characterize the entities present in an image from different an gles and, therefore, can be analyzed by means of multivariate techniques to assist in image enhancement and classification. The user-friendly program prepared can be used in conjunction with GIS (Geographic Information System ) software such as ArcView to implement the spatial and scaling method. it has been applied to the stream sediment geochemical data set (923 grid samp les) for gold mineral exploration in the Habahe map sheet, Altay Shan, Xinj iang, northwestern China. The spatial and scaling method provides better re sults than the ordinary moving average method. (C) 1999 Elsevier Science B. V. All rights reserved.