Spatial and scaling modelling for geochemical anomaly separation

A spatial and scaling approach with a user-friendly windows program is introduced which can be used to assist exploration geologists and geochemists in geochemical data analysis and anomaly separation. It can also be used for image enhancement and classification. Statistics are calculated and optimized within a variable-sized moving window centred at an arbitrary sample location. The moving window has both variable size and shape determined by three parameters: r (size), β (ratio of long and short axes), and θ (orientation). It calculates five optimal indexes for each sample location: the optimal statistic U(r₀,β₀,θ₀), optimal size r₀, shape indexes β₀ and θ₀, and scaling index α (singularity exponent). These indexes characterize the entities present in an image from different angles 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 samples) for gold mineral exploration in the Habahe map sheet, Altay Shan, Xinjiang, northwestern China. The spatial and scaling method provides better results than the ordinary moving average method.