Statistical evaluation of multiple interpolation techniques for spatial mapping of highly variable geotechnical facets of soil in natural deposition

Zain Ijaz, Cheng Zhao, Nauman Ijaz, Zia Ur Rehman, Aashan Ijaz

Research output: Contribution to journalArticlepeer-review

Abstract

This article aims to build geotechnical maps (GMs) of soil in its natural deposition using the most efficient interpolation techniques based on statistical analyses using a comprehensive field and laboratory-investigated soil database. GMs were developed based on standard penetration test (SPT-N), shear wave velocity (Vs), plasticity index (P.I), activity (A), linear shrinkage (L.S), cohesion (c), frictional angle (Φ), and chemical contents in conjunction with the recommendation for foundation design. The performance of various interpolation models was evaluated using key performance indices (KPIs) revealing that inverse distance weighting (IDW) interpolation based on the modified shepherd method demonstrated a significant validation with the least prediction error and thus opted for the development of GMs. The created GMs reveal the predomination of fine-grained soil at shallow layers (1.5–3.0 m) with SPT-N and Vs ranging between 0–20 and 0–300 (m/s); P.I between 0- < 17; A-value between 0- < 1.25; L.S between 0- < 7; c-value between 0-3psi; and Φ between 28–30°. The deeper stratum beyond m 3.0 m depth is dominated by coarse-grained soil, cobbles, and boulders. Moreover, the chemical contents at 1.5 m strata fall under the permissible thresholds. For the integrated GMs at multiple strata, the efficacy of GMs was quantified in terms of the root means square error (RMSE) and mean absolute error (MAE) that vary from 0.14 to 1.98, while the Nash Sutcliffe model efficiency coefficient (NSE) and Pearson coefficient (PC) ranged between 0.83–0.99. In addition, the pertinent recommendation maps for foundation design were developed and problematic areas were highlighted that need due attention.
Original languageEnglish
Pages (from-to)105–129
JournalEarth Science Informatics
Volume16
Early online date5 Jan 2023
DOIs
Publication statusPublished - 1 Mar 2023

Keywords

  • Soil investigation
  • Spatial mapping
  • Geotechnical database
  • GIS
  • Statistical evaluation
  • Key performance indicators

Cite this