Accurate seabed substrate mapping is essential for marine spatial planning, benthic habitat characterization, and infrastructure routing, yet traditional ground-truthing methods remain costly and spatially limited. We demonstrate a cost-effective workflow for regional-scale substrate mapping using single-beam echo sounder (SBES) data on the northwestern Black Sea shelf. During a June 2019 survey across three polygons (~700 km²) adjacent to major Ukrainian ports, we collected approximately 230,000 acoustic pings and 22 ground-truth sediment samples. We extracted RoxAnn-type energy metrics (E1: roughness proxy; E2: hardness proxy) and applied unsupervised hierarchical clustering to delineate six acoustic substrate classes: soft muds, medium muds, hard muds, silty shells, sandy shells, and shelly sands. The classification achieved an overall mean silhouette coefficient of 0.46, with end-member classes (soft muds, shelly sands) showing strong separation (silhouette >0.74) and transitional classes reflecting natural environmental gradients (silhouette 0.39-0.46). Multidimensional scaling revealed clear separation along gradients of consolidation state and bioclastic content. Ground-truth validation confirmed strong correspondence between acoustic signatures and field lithology, with end-member classes showing high silhouette values (>0.74) and transitional classes reflecting natural environmental gradients. Depth-dependent probability modeling enabled extrapolation of substrate classifications across unsurveyed areas, exploiting systematic depth stratification driven by wave-energy gradients and riverine sediment deposition patterns. Comparison with EMODnet seabed layers showed that the SBES workflow resolves finer-scale heterogeneity and more detailed transitional zones than legacy point-based compilations. Our acoustic classification provides operationally relevant information on consolidation state and shell content that granulometric schemes alone cannot capture, with direct implications for cable burial feasibility, benthic habitat assessment, and dredging operations. Recent disturbances including the 2023 Kakhovka Dam breach and 2024 Kerch Strait oil spill underscore the need for repeatable, rapid substrate mapping to track sediment redistribution and contamination. This study demonstrates that low-cost SBES systems, coupled with transparent statistical workflows and minimal ground-truth calibration, can produce credible substrate maps at regional scales, supporting improved baseline characterization and time-series monitoring for under-surveyed shelf regions.
Goncharov et al. (Thu,) studied this question.