AUV-Based Small Scale Quantitative Predictive Mapping of Mn-Nodules: Precondition for Ideal Mining Path Planning

"Ferromanganese nodules are again of interest because of their high concentrations of Ni, Co, Cu as well as REE. With regards to an operational underwater mining industry, one of the main questions and important requests is to finding a method that can predict the abundance of nodules with high accuracy within reasonable time spans, without many requirements for ground truth data; and all these on an operational scale and ideally automated. In this regard, we present results of a study based on the combination of highresolution multibeam and optical data acquired by an AUV using specific machine learning techniques to reveal a detailed distribution of the Mn-nodules in correlation to the seafloor terrain. The accuracy of the method employed and results gained were validated and approved through comprehensive statistical analysis and by comparison to box-core data.Study AreaOur study area (AUV survey block G77) is located in the Clarion-Clipperton Zone (CCZ) (Eastern Central Pacific Ocean), an area that has been receiving international research and industrial attention for many years already (e.g. McKelvey et al., 1979; Bernhard and Blissenbach, 1988, von Stackelberg and Beiersdorf, 1991, Jeong et al. 1996). Block G77 covers an area of 9.5km2 in a depth between -4478m and -4536m. The area is characterised by gentle slopes 0-5° (0-3° in most of the area and only in the outer parts of the Block slopes can exceed 5°). The seafloor is characterised by moderate to low rugosity, the only exception being a small area in the eastern part, where sub-recent smallscale volcanic activity created 15 cone-shaped morphological features of up to 55m height and 150m width that are clustered in an area of 700 x 380m (Figure 1)."