A data-driven approach to understanding the relations between geothermal exploration parameters: insights from Coso, Brady and Desert Peak, USA

Understanding the relations between indicator minerals, faults and temperatures is crucial for effective geothermal exploration. This study employs a novel data-driven analysis to investigate these relations across three distinct geothermal fields in California and Nevada in the USA: Coso, Brady and Desert Peak. Utilizing fault maps, land surface temperatures (LSTs) and indicator mineral maps (alunite, chalcedony, hematite, kaolinite and opal), non-spatial analyses (Kolmogorov–Smirnov tests, ANOVA and linear regression) were combined with spatial analysis (co-location analysis) to uncover associated patterns. At Coso and Brady, alunite and kaolinite are closely linked to faults and elevated LSTs, while chalcedony shows negative correlations. Hematite aligns positively with LSTs but negatively with faults, and opal correlates positively with faults but not with LSTs. These trends align with geothermal alteration models that involve sulfuric acid or alkali-chloride fluid discharges. Desert Peak's blind geothermal system contrasts with the negative relations between its indicator minerals, faults and LSTs. This research provides insights into integrating diverse datasets to enhance exploration strategies and reduce uncertainty. Fault surveys are essential for identifying heat and fluid pathways, while LST surveys are critical for detecting blind systems. In addition, the study underscores the need for data-driven approaches to explore subtle geological characteristics in geothermal exploration.