Integrated quantitative interpretation of multiple geophysical data for geology differentiation

The future of mineral exploration depends on innovative methods of data integration and interpretation because new discoveries are becoming scarcer over the years. As brownfield exploration areas reach maturity and greenfield exploration faces increasingly deeper targets and targets hidden under cover, geophysics is becoming the primary exploration tool. When little a priori geological information is available, such as in greenfield exploration, multiple geophysical methods are necessary to improve interpretation and decrease exploration risk. However, it is challenging to deal with multiple geophysical methods in geologically complex areas. For this reason the main motivation of my thesis is to develop integrated quantitative interpretation methods of multiple geophysical data for geology differentiation. Multiphysics is fundamental for identifying geological units instead of just identifying isolated geophysical anomalies in different physical property models. It also allows uncertainties to be minimized if all the available data are properly integrated. Therefore, I first develop a method for geology differentiation based on spatially limited geological information and general relations of physical properties that can be applied to geophysical data over a large area. Then, in the absence of geological information, I incorporate more geophysical data and develop a method of geology differentiation by applying unsupervised machine learning (correlation-based clustering) for the construction of a quasi-geology model. Additionally, I develop a novel method to improve the construction of susceptibility models, the magnetic on-time transient electromagnetic (MoTEM) method. The use of more accurate physical property models improve the geology differentiation. The research I have developed contributes to solving practical challenges of greenfield mineral exploration by providing effective unbiased integrated interpretation methods that produce directly interpretable quasi-geology models.

Name thesis:degree_name

Doctor of Philosophy (Ph.D.)

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Geophysics

Grantor dc:publisher

Colorado School of Mines. Arthur Lakes Library

Year dc:date.issued

2018

Author dc:creator

• Melo, Aline Tavares

Advisor dc:contributor.advisor

• Li, Yaoguo

Committee members dc:contributor.committeemember

• Hitzman, Murray Walter
• Dagdelen, Kadri
• Sava, Paul C.
• Swidinsky, Andrei

• IOCG
• mining
• machine learning
• inversion

Statement dc:rights

• Copyright of the original work is retained by the author.

Language dc:language.iso

eng, English

Identifier

T 8497

OAI identifier oai:identifier

oai:repository.mines.edu:11124/172327