Advanced geospatial modeling and statistical evaluation of heterogenous geotechnical facets at multiple depths using the improved formulation of modified Shepard IDW algorithm

A formidable challenge in geology and geotechnics is the significant spatial heterogeneity in the subsoil characteristics in fine-scale grids. To mitigate this problem, geotechnical data is integrated as geotechnical soil maps (GSMs), which uses sophisticated interpolation techniques and provides an advanced understanding and accurate depiction of subsurface variability. This study uses an improved formulation of inverse distance weighting (IDW) algorithm based on modified Shepard method, integrated with the Google Earth Engine platform. The prediction efficiencies of GSMs and traditional IDW algorithm are statistically evaluated and compared, considering heterogeneous geotechnical facets at multiple depths in an unexplored region. Pertinent geotechnical properties including soil type, plasticity index, and standard penetration test were considered to evaluate the algorithm performance based on critical performance metrics. The results demonstrate that the improved formulation of the IDW algorithm is more relevant to field values and tends to align with Tobler's first law of geography by inducing a smooth transition rather than a disruptive trend owing to high geotechnical variability. The prediction accuracy increased by 10 - 20% compared to the traditional IDW algorithm.