Quantifying uncertainty and improving prospectivity mapping in mineral belts using transfer learning and Random Forest: A case study of copper mineralization in the Superior Craton Province, Quebec, Canada

• Published in: Ore geology reviews Vol. 166; p. 105918
• Main Authors: Lauzon, Dany, Gloaguen, Erwan
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2024
• Subjects: Data integration; Mineral belts in the Superior Craton Province; Mineral prospectivity mapping; Quantification of uncertainty; Transfer learning; Mineral prospectivity mapping; Mineral belts in the Superior Craton Province; Transfer learning; Quantification of uncertainty; Data integration
• ISSN: 0169-1368; 1872-7360
• DOI: 10.1016/j.oregeorev.2024.105918

Left) Geological provinces of Quebec, Canada, and subprovinces of the Superior Province (pink units). The labelled black rectangles show the five study areas within the Superior Province. Right) Regional geological maps, fault locations (black lines), and copper occurrences (orange symbols) of the studied areas. Each area is divided into two (black dotted lines), one for training (a) and one for validation (b). Mineral belts are associated with metavolcanic and metasedimentary rocks, labelled in shades of green. The complete color codes can be consulted at the following address: MERN (2014). (I- Matagami region, II- Chibougamau area, III- Rouyn-Noranda District, IV- Troilus region, and V- North zone of the La Grande subprovince). Figure adapted from SIGÉOM (2022). [Display omitted] •• A transfer learning methodology was proposed for mineral prospectivity mapping.•• Random Forest Regressor was employed to produce geochemical maps consistent with geology.•• The proposed methodology was able to predict more accurately hotspots for exploration targeting.•• The transfer learning approach is applicable to any areas with similar geological, and deposit context.•• The algorithm was tested and validated using five mineral belts in the Superior Craton Province, Quebec, Canada. Mineral prospectivity mapping (MPM) involves identifying locations with a higher potential for mineral exploration based on a set of explanatory variables. In cases where there is a scarcity or absence of unfavorable sites that adequately represent the geological context for deposit discovery, generating synthetic negative data sets becomes necessary to employ a machine learning algorithm optimally. Moreover, when favorable sites are insufficient for deposit discovery within a geological zone, machine learning methods can potentially result in large and highly uncertain prospecting areas. This article proposed a concept based on transfer learning by applying the knowledge gained from mineral belt signatures in different geological zones to a related area. The positive training data were taken from five mineral belts distanced from each other, while the negative data were sampled using geological constraints based on the distance to occurrences and spatial associativity. The results demonstrate that transfer learning, combined with geological constraints applied to the creation of negative datasets, improves model performance and prediction of known deposits while significantly reducing uncertainties. Mineral prospectivity models for predicting potential copper formations were generated using data from the Quebec Government's spatial reference geomining information system, SIGEOM. The case study for this work focused on the geological province of the Superior Craton, which encompasses the vast majority of northeastern Quebec.