Key trace element mapping for mineral prospectivity
This exciting project has generated a long-term research collaboration between a local business, Moon Geology, and Deep Digital Cornwall (DDC) staff at the Camborne School of Mines (Robin Shail, Nick Harper, Luke Palmer) and Cornish Lithium (Chris Yeomans, Fred Jackson).
Exploratory conversations between Moon Geology and the DDC team commenced in early 2021 and resulted in an application to the DDC Grant Fund for a £5k Innovation Voucher. Moon Geology used this funding to analyse 72 Cornubian granite samples from across Devon, Cornwall and the Isles of Scilly for Fluorine (F), Chlorine (Cl) and Boron (B) data.
Advanced data reduction methods and supervised machine learning techniques were used to investigate further the Simons et al. (2016) granite classification scheme. The team discovered that the granite types can be better separated as distinct geochemical classes, further reinforcing the granite classification scheme. Supervised machine learning methods, such as random forests and support vector machines, suggest that predicting these classes should be effective when trialled on future datasets.
The project outputs are expected to include publications in high-impact journals during 2022 authored jointly by Moon Geology and the DDC team. The data and results from the machine learning analyses will be made available for exploration within the DDC Data Hub and Visualisation Suite.
