Multispectral discrimination of spectrally similar hydrothermal minerals in mafic crust: A 5000 km2 ASTER alteration map of the Oman–UAE ophiolite

Multispectral remote sensing of hydrothermal alteration in volcanogenic massive sulfide (VMS) ore systems in mafic crust is relatively uncommon, in part due to the short-wave infrared spectral similarity of several key alteration minerals: epidote, chlorite, actinolite, and serpentine. In this study, we developed regional mosaic generation and classification workflows for Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) imagery to discriminate these minerals over the entire crust of the Semail ophiolite (Oman–UAE). Spectral discrimination was achieved through adaptation of the ASTER (pre-)processing workflow to the specific mapping targets, available datasets, and location of this study. Necessary steps included the pre-selection of ASTER scenes without residual atmospheric water features, mosaic normalization based solely on overlapping target outcrops, correcting cross-mosaic ramp errors, and alteration map classification based on image-derived reference data. The resulting alteration map, validated through comparison with field mapping and sampling, is the most areally extensive continuous survey of hydrothermal alteration yet presented for oceanic crust, providing a renewed framework for research and mineral exploration of Earth's largest ophiolite. Our map confirms that the vast majority of the upper oceanic crust is regionally altered to a spilite type secondary mineral assemblage. Localized areas of epidosite alteration, marking focused hydrothermal flow paths, are confined to the upper oceanic crust, whereas areas of previously unrecognized but intense actinolite alteration are common in both the lower and upper oceanic crust. Our methodological developments expand the standard considerations necessary for regional geological mapping using infrared image mosaics. They further demonstrate the underappreciated capability of multispectral data for mapping spectrally similar rock types. Although the specifics of the method are necessarily optimized for the Oman–UAE ophiolite, re-optimization based on local reference data should allow similar results to be achieved in other well-exposed mafic-hosted VMS districts.