Our recent study highlights the critical role of discovering new copper deposits for advancing renewable energy and zero-emissions transport, focusing on the connection between porphyry systems enriched in copper and the evolution of subduction zones. Utilising advanced machine learning within a spatiotemporal framework, we analyse 47 Cenozoic intrusion-related copper-gold deposits in Papua New Guinea and the Solomon Islands, developing a model that accurately predicts mineral occurrences and pinpoints potential mineralisation indicators. It finds that subduction zone obliquity angle, arc curvature and length, rapid plate convergence, and specific seafloor spreading rates significantly influence mineralisation. This novel approach, blending plate motion models with machine learning, offers new insights into porphyry ore formation and proposes enhanced exploration criteria for identifying copper deposits in both active and extinct subduction zones.
- pygplates
- gplately
- cartopy
- geopandas
- matplotlib
- netCDF4
- numpy
- pandas
- pulearn
- scipy
- shapely
- skimage
- sklearn
- skopt
- tensorflow
@article{Farahbakhsh2024,
title = {Advanced machine learning-based spatiotemporal prospectivity modelling of porphyry systems in Papua New Guinea and the Solomon Islands region},
author = {Farahbakhsh, Ehsan and Zahirovic, Sabin and McInnes, Brent I. A. and Polanco, Sara and Kohlmann, Fabian and Seton, Maria and M{\"u}ller, R. Dietmar},
year = {2024},
journal = {?},
volume = {?},
number = {?},
pages = {?},
doi = {?},
}