Lithological mapping of Waiotapu Geothermal Field (New Zealand) using hyperspectral and thermal remote sensing and ground exploration techniques

•Lithology mapping using hyperspectral data combined with machine learning classification.•Elevated Ag, Au, S and Sb of hand samples correlate to waters of parental origin.•Temperature analysis per lithology by combining TIR and hyperspectral imagery. Hydrothermal systems develop secondary mineral assemblages due to rock and fluid interactions. Such minerals can be identified on the surface and mapped using hyperspectral imaging and thermal remote sensing to complement geothermal exploration and monitoring. These methods are effective in detecting hydrothermal alteration and thermal anomalies of extensive areas, faster and cheaper than ground-based mapping alone. Here, we study the Waiotapu Geothermal Field, located in the Taupo Volcanic Zone, New Zealand, to demonstrate the capability of such remote sensing tools for mapping surface hydrothermal lithologies and geothermal features. Application of remote sensing techniques to study the Waiotapu Geothermal Field is an optimal alternative as the site is protected for its sinter deposition springs, geysers, vulnerable geothermal features and rare thermotolerant vegetation. Airborne hyperspectral and thermal remote sensing data were complemented with 75 soil and rock samples for ground truth. These samples have been analysed using Visible Near Infrared (VNIR) and Shortwave Infrared (SWIR) spectroscopy, Scanning Electron Microscopy (SEM) with Energy Dispersive X-ray Spectroscopy (EDS) and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) to constrain alteration mineralogy and bulk rock geochemistry. Lithological maps created through Spectral Angle Mapper (SAM) and Support Vector Machine (SVM) supervised classifications using the airborne hyperspectral data benchmarked against independent image- and ground-based validation data. Three main lithologies were mapped out, including silica sinter deposits, mixed alteration zones, and acid-sulphate alteration at 57.3% for SAM and 87.5% for SVM overall accuracy. The lithological classes have been analysed using the thermal imagery, which confirmed strong spatial heterogeneity of surface heat sources and overall higher surface temperatures for acid-sulphate alteration zones. Through Principal Component Analysis, enrichment of Ag, Au, S and Sb are found to be located around the Champagne Pool, which is where the near-neutral pH high-chloride fluids discharge at surface. The combination of airborne hyperspectral and thermal imagery with ground-based methods applied at