High-resolution Digital Elevation Models (DEMs) are critical for accurate coastal flood risk assessment, particularly in low-lying areas vulnerable to sea-level rise, storm surges, and other natural hazards. This study presents a comparative evaluation of five DEMs with varying spatial resolutions and data sources for a test area along the northeastern Bulgarian Black Sea Coast around the Durankulak Lake—an area identified as at risk of coastal flooding. A high-resolution Uncrewed Aerial Vehicle (UAV) derived DEM was generated using Structure-from-Motion and Multi-View Stereo (SfM-MVS) photogram-metry and compared against three other DEMs: BG DSM (4 m), MGS DEM (10 m), and re-sampled TanDEM-X (10 m). The vertical accuracy of DEMs was assessed using control points (CPs), including GNSS-measured and geodetic reference points, and parametric and non-parametric statistical methods, along with statistical tests to assess the error distribution. Results indicate that the UAV and BG DSM models exhibited normal error distributions with high vertical accuracy (root mean square error: 0.098 m and 0.776 m, respectively), while the MGS and TanDEM-X models display significant deviations from normality and lower precision. The UAV-derived DEM has offered the highest accuracy and reliability, though it requires significant field and processing resources. The findings highlight the importance of selecting DEMs based on the required precision, available resources, and application-specific context, particularly for hydrologic and hydrodynamic modeling and flood risk mapping under the EU Floods Directive. This study provides a robust methodological framework for evaluating DEM quality and suitability in coastal hazard assessments and contributes to improved geospatial data practices in flood-prone regions.

Coastal flooding, coastal zone, Durankulak Lake, natural hazards, UAV photogrammetry, 3D terrain model