The global shift toward sustainable energy is driving an unprecedented Cu demand, for which porphyry Cu systems are pivotal. Although these deposits are linked to prolonged compressional periods, the mechanisms controlling their precise spatiotemporal distribution remain unclear. To address this, we employed a multidisciplinary approach combining machine learning and geological analysis to evaluate the best parameter combination for Neogene Andean porphyry Cu system formation (0°S−38°S). The models included subduction parameters—margin-normal convergence rate, obliquity of subduction, and subducting ocean floor age—as well as factors related to aseismic ridge subduction, rarely considered in similar studies. To characterize aseismic ridges, we designed a parameter termed “migration rate,” which represents a proxy for their along-trench movement over time. This parameter was handled by first identifying the nearest ridge to each deposit position and incorporating it into the models as an interaction term, calculated by multiplying it with the corresponding ridge-deposit distance (migration rate × distance, or MDI). The results reveal a strong correlation between aseismic ridge migration patterns and metallogenesis, given that MDI was the most impactful parameter in the models, followed by obliquity of subduction. Lower MDI values positively impacted the predictions, suggesting that high-Cu (2 Mt) deposits preferentially formed in proximity to nonmigratory ridges. Based on these findings, we propose that nonmigratory ridges produced more efficient local compression, ultimately leading to better ore-forming conditions in the overriding plate. Furthermore, the models reasonably predicted the high-Cu regions within the area, demonstrating their potential for exploration in applications such as space-time probability mapping.
Castillo et al. (Tue,) studied this question.