Subsurface Topographic Modeling Using Geospatial and Data Driven Algorithm

Abstract

Infrastructures play an important role in urbanization and economic activities but are vulnerable. Due to unavailability of accurate subsurface infrastructure maps, ensuring the sustainability and resilience often are poorly recognized. In the current paper a 3D topographical predictive model using distributed geospatial data incorporated with evolutionary gene expression programming (GEP) was developed and applied on a concrete-face rockfill dam (CFRD) in Guilan province- northern to generate spatial variation of the subsurface bedrock topography. The compared proficiency of the GEP model with geostatistical ordinary kriging (OK) using different analytical indexes showed 82.53% accuracy performance and 9.61% improvement in precisely labeled data. The achievements imply that the retrieved GEP model efficiently can provide accurate enough prediction and consequently meliorate the visualization insights linking the natural and engineering concerns. Accordingly, the generated subsurface bedrock model dedicates great information on stability of structures and hydrogeological properties, thus adopting appropriate foundations.