AUV hydro-acoustic and optical data enhance high resolution quantitative mapping of deep sea hard substrates such Mn-nodules. Machine learning algorithms predict with good accuracy the Mn-nodules abundances over large scale areas utilizing one third of ground truth optical data. Accurate maps of Mn-nodule abundances raise new questions about the role of fine scale geomorphology in nodule formation, provide new insights in deep sea ecological studies, and improve mineral assessment estimations.