Application of very high-resolution satellite imagery to identify silica-rich rock for future cosmogenic exposure dating in remote unvisited areas of Antarctica.

Abstract

Rock outcrops protruding above the ice surface in Antarctica (nunataks) can provide direct geologic evidence for past ice sheet fluctuations through the measurement of concentrations of cosmogenic nuclides that accumulate in their surfaces once the rock is exposed. Felsic lithologies, which are typically pale in colour and dominated by quartz, feldspars, and micas, are suitable for exposure age dating since quartz is the often-preferred target mineral for extraction of the rare cosmogenic isotopes which make deglacial reconstructions possible. The geology of rock outcrops in Antarctica are, however, often sparsely mapped and many exposures are challenging to access due to both their remoteness and the extreme conditions typically encountered on the continent. Satellite based spectral mapping offers an effective way to characterise the geology of large areas of exposed rock rapidly and safely in regions where it is logistically very challenging and expensive to conduct fieldwork. Remote sensing therefore offers a valuable method for preliminary characterisation of an area’s suitability for eventual targeted retrieval of cosmogenic nuclide samples.   Previous studies found that the Thermal Infra-Red (TIR) sensor onboard the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) is very effective at discriminating rock types by their silica content, but spectral mapping of smaller felsic rock outcrops in Antarctica has been constrained by its low spatial resolution (90 m). Here we assess the potential of multispectral remote sensing using both ASTER and very high-resolution Worldview-3 (WV-3) imagery to distinguish felsic from mafic rock outcrops at visible-near infrared (VNIR) and shortwave infrared (SWIR) wavelengths. At Mount Murphy, a remote site in West Antarctica more than 1,600 kilometres from both the US Antarctic Program’s McMurdo Station and the British Antarctic Survey’s Rothera Research Station, we identify four dominant rock types from our spectral mapping: granites, gneisses, basalt and fragmental hydrovolcanic rocks (hyaloclastite). Image derived spectral profiles of these four rock types were used as input for spectral classification and lithological mapping of the Mount Murphy site. Supervised classification results indicate that WV-3 performs well at differentiating felsic from mafic rock types and that ASTER imagery, while coarser in resolution, can also achieve satisfactory results, and could therefore be used in concert with more targeted WV-3 image acquisitions. We also demonstrate that separation of mafic (fragmental) hydrovolcanic and basalt rock types can be achieved at VNIR-SWIR wavelengths, a result that will be useful for future spectral mapping of volcanic rocks on other planets. We used spectral mapping and supervised classification results to produce a new geologic map of Mt Murphy. Overall, our results demonstrate the potential of spectral mapping and classification using WV-3 and ASTER datasets to identify and characterise suitable sites for future cosmogenic nuclide sampling campaigns.