GEOSCAN Search Results: Fastlink


TitleExploratory spatial modelling; demonstration for Carlin-type deposits, central Nevada, USA, using Arc-SDM
AuthorRaines, G L; Bonham-Carter, G F
SourceGIS for the earth sciences; by Harris, J R (ed.); Geological Association of Canada, Special Paper 44, 2006 p. 23-52
LinksOnline - En ligne
Alt SeriesEarth Sciences Sector, Contribution Series 20070163
PublisherGeological Association of Canada (St. John's, NL, Canada)
Mediapaper; CD-ROM
RelatedThis publication is contained in Harris, J R; (2006). GIS for the earth sciences, Geological Association of Canada, Special Paper vol. 44
File formatpdf
AreaCentral Nevada; United States of America
Lat/Long WENS-119.0000 -116.0000 41.0000 38.5000
Subjectsmathematical and computational geology; mineral exploration; mineral potential; models; modelling, structural; digital terrain modelling; remote sensing; data collections; computer simulations; computer applications; computer mapping
Illustrationslocation maps; tables; bar graphs; schematic diagrams; graphs; digital images
AbstractArc-SDM (Arc-Spatial Data Modeller) is a software extension to ArcView GIS that is useful for mineral exploration and other types of spatial prediction using regional datasets. This paper is aimed mainly at those learning to use Arc-SDM, and may be used in combination with the program documentation and various applications papers. Regional datasets related to Carlin-type deposits in Nevada are used throughout to illustrate the methodology.
Spatial evidence from regional geological and geochemical datasets is used to generate maps showing prospectivity for Carlin deposits in the north-central Nevada in the Basin and Range province of the western United States. The purpose of the paper is to illustrate how the Arc-SDM extension to ArcView 3 can be used for this type of modelling. The approach is not a 'black box' computer operation, with input data fed into it at one end and a prediction map coming out at the other. Instead, a series of subjective decisions are required, strongly dependent on an understanding of the characteristics of the mineral deposits and their exploration guidelines, augmented by statistical analysis of the spatial associations between known mineral sites and the GIS data layers (evidential themes). The data used in the example comprise the geological map, a map of faults, and a regional stream sediment geochemical survey.
This paper will be valuable for those learning to apply Arc-SDM, because the reasons for making decisions about reclassification to binary or multiclass themes, for example, are discussed using the weight tables generated for the geological, fault, and geochemical maps. Five models and two sets of neural network models are created. Model 1 uses binary reclassifications only and applies weights of evidence and weighted logistic regression. Model 2 uses multiclass reclassifications with the same methods. Model 3 is similar to Model 2 except that a principal components analysis of the geochemical data is employed. Model 4 is a fuzzy logic model using both the fuzzy AND and fuzzy GAMMA operators to combine fuzzified evidence. Model 5 is similar to Model 4 except only the fuzzy AND is applied. The neural network set of models are created with a radial basis function link net and a fuzzy neural network in order to demonstrate the complementary nature of these modelling approaches to the other methods.
Modelling using any or all of these methods is valuable for exploring spatial associations in the data, for making predictions, for obtaining estimates about uncertainty of prediction, for extracting and ranking spatial evidence by its utility to predict deposits, for identifying data gaps, for testing hypotheses, for integrating diverse datasets, and for communicating the results at various levels of expertise. In short, the Arc-SDM extension provides a powerful set of spatial modelling tools for exploring spatial data and making predictions. An important conclusion to be drawn from the use of Arc-SDM is that although a set of guidelines can be followed, there is no "right" way to do things - the process of using spatial data for understanding and prediction is still an exercise that requires a good understanding of both the data and the mechanisms that may have controlled or influenced their generation, and there are subjective decisions to be made at many stages