Experimental Application of Underwater Hyperspectral Imaging on Seafloor Massive Sulphides from the Loki’s Castle on the Arctic Mid-Ocean Ridge, Norway

INTRODUCTION Efficient exploration methodologies for base metal deposits have huge benefits for future deep sea mining endeavours, and underwater hyperspectral imaging (UHI) has been variably demonstrated to have applications in the identification and characterisation of mineralisation on the seafloor for both nodules and seafloor massive sulphides (SMS) (Dumke et al., 2018 and 2017 respectively). Due to growing resource requirements (Singer, 2017), SMS deposits are considered to be increasingly significant sources of copper and zinc, as well as silver and gold. This style of mineralisation is currently occurring at active hydrothermal vents (black and white smokers) but now-inactive sites also have the potential to be large mineral deposits. As such, the development of novel methodologies to detect both on-going and extinct mineralisations are of high importance. Geological setting During the MarMine cruise (Ludvigsen et al., 2016), non-mineralised host rock, low grade ore and high grade ore grab samples (Snook et al., 2018) was collected by ROV from the Loki’s Castle deposit (Pedersen et al., 2010), located on the Mohn’s/Knipovich junction on the ultra-slow spreading Arctic Mid-Ocean Ridge (AMOR), at a depth of approximately 2300 m. This material has been characterised as part of the MarMine project (e.g. Snook et al., 2017), and, by imaging compositionally constrained material, the measured hyperspectral responses can be used to generate a library for identification of deposits on the ocean floor. UHI technology Remote sensing with hyperspectral and multispectral technologies has seen wide use in prospecting for ores and hydrocarbons on land (Van der Meer et al., 2012). Iron oxides and sulphides have low blue reflectance and high red reflectance in the visible spectrum. The ratio between these two pairs has been used in exploration for terrestrial deposits of hydrothermal origin (Sabins, 1999). This discrimination typically also includes wavelengths exceeding the visible spectrum (1.5µm – 2.5µm). These wavelengths are typically not utilised underwater because infrared wavelengths and higher are rapidly attenuated in water. However, discrimination based on the full shape of the spectral response in the visible light (350nm – 800nm) may still be possible (Bolin and Moon, 2003). Resolving the endmembers from the spectral responses requires prior knowledge about the optical properties of the materials, i.e. their reflectance. These reflectances can be obtained in a controlled environment, where the spectra of the lamps, wavelength- dependent attenuation in water and scattering of light can be accounted for (Johnsen, 2013).