Geospatial-based approach for susceptibility assessment of expansive soils using a new multicriteria classification model

Geotechnical spatial distribution was applied using the geostatistical interpolation approach to improve the accuracy of spatial prediction of geotechnical data throughout the study area and characterize the expansive soil for highways and structure projects. The cross-validation method was used to calibrate the digital models (DM) and to determine their level of accuracy, during the process of establishing the optimal geotechnical classification strategy through 187 samples, including their water content (w), dry bulk density (γd), plasticity index (PI), methylene blue value (MBV), carbonate content (CaCO3), and free swelling pressure (Ps). The spatial interpolation was performed by applying sequential spherical, Gaussian, and exponential simulations. Following calibration, a new multicriteria soil classification (MSC) formula was developed using a statistical approach to estimate the classes of expansive soil susceptibility indirectly. The susceptibility (DM) indicated a high correlation coefficient (R2) of 0.99. A susceptibility map was created based on the new developed formula, and it revealed that the study area could be divided into four principal parts: extremely susceptible, susceptible, moderate, and safe. The swelling soil associated with prospective pavements and building damage can be estimated based on spatial zoning maps considering geotechnical and geological parameters and comparing the spatial correlations of geotechnical site classes. These maps could develop soil improvement strategies in civil and geotechnical engineering fields.