Density-adaptive geospatial framework for high-resolution seismic microzonation in digital construction

Abstract

This study presents an advanced geospatial framework for Seoul, utilizing a database of over 29,000 boreholes to develop a high-resolution pseudo-3D shear-wave velocity field. By implementing site-specific SPT-N correlations, the framework captures intra-stratal heterogeneity and vertical stiffness gradients often masked by stratigraphic averaging. A density-adaptive mosaic-based interpolation was employed to estimate bedrock depth (H), time-averaged shear wave velocity (V_S30) and fundamental period (T_G), effectively mitigating the “smoothing effect” inherent in ordinary kriging. The model achieved an average 36.56% reduction in cross-validation RMSE across all primary parameters and a 50% reduction in V_S30 RMSE against 9999 independent downhole records. T_G was identified as the critical factor for localized resonance risk. The resulting 5 m resolution zonation maps delineate seismic hotspots and alluvial micro-zones previously obscured in macro-scale assessments, providing a robust geospatial template for lot-level hazard mitigation and seismic design code refinement within the digital construction paradigm.