| Title | Mineral potential mapping: examples from the Red Lake greenstone belt, northwest Ontario |
| |
| Author | Harris, J R; Sanborn-Barrie, M |
| Source | GIS for the earth sciences; by Harris, J R (ed.); Geological Association of Canada, Special Paper 44, 2006 p. 1-21  Open Access |
| Links | Abstract and online ordering / Résumé et commande en-ligne
|
| Image |  |
| Year | 2006 |
| Alt Series | Geological Survey of Canada, Contribution Series 2005076 |
| Publisher | Geological Association of Canada (St. John's, NL, Canada) |
| Document | serial |
| Lang. | English |
| Media | paper; CD-ROM |
| Related | This publication is contained in GIS for the earth sciences
|
| File format | pdf |
| Province | Ontario |
| NTS | 52N/04 |
| Area | Red Lake |
| Lat/Long WENS | -94.0000 -93.5000 51.2500 51.0000 |
| Subjects | metallic minerals; structural geology; mapping techniques; computer mapping; gold; greenstone belts; mining; mines; mineral exploration; mineral potential; sulphide deposits; modelling,
structural |
| Illustrations | tables; location maps; geological sketch maps; graphs |
| Program | Northern Resources Development Program |
| Released | 2006 01 01 |
| Abstract | Geographic Information Systems (GIS) are now commonly used in concert with various spatial modelling and statistical software packages by exploration companies to generate maps to assist in targeting
various mineral commodities. A number of different modelling methods may contribute to the generation of such mineral potential maps. This paper provides a general review of the mineral potential mapping process using a GIS-based system and
highlights a number of key issues that are pertinent to the production of prospectivity maps. These include the effect of using sparse vs. more extensive datasets, the choice of binary vs. continuous surface format for input data, the impact of
utilizing different modelling algorithms including weights of evidence, weighted logistic regression, fuzzy logic and neural networks, and the consequences of using different populations of gold occurrences (e.g., mines vs. prospects vs. showings) as
control populations ("training sets") for modelling. The impact of these issues on the GIS mineral potential mapping process is explored using data from the Red Lake greenstone belt of Ontario, Canada, one of Canada's most prolific gold districts.
The most reliable gold potential maps generated for the Red Lake greenstone belt were produced from datasets that included many evidence maps rather than just a few evidence layers, demonstrating that the use of sparse datasets is likely to adversely
affect the degree to which the derived mineral potential maps identify prospective areas and predict known mineral deposits. The use of evidence maps in binary format resulted in gold potential maps that were slightly better predictors than maps
produced using continuous surface evidence maps. However, the results vary depending on how thresholds (i.e., geochemical concentration, proximity to a contact, etc.) are determined when creating the binary map. The different data-driven modelling
methods investigated in this study resulted in gold potential maps that are similar with respect to the areas identified as high potential. This indicates that the input data (evidence maps) used for a particular exploration model are probably more
critical than the modelling method chosen to combine the data. The potential maps derived for the Red Lake greenstone belt identified several new areas with geological and/or geochemical characteristics that are similar to the producing mines; these
new areas represent prospective targets for further exploration. |
| GEOSCAN ID | 220633 |
|
|