Evaluation of Spatial Variability of Loess Properties Using a Self-Developed Database in Shaanxi Province, China

Loess is a natural Quaternary aeolian sediment, which is widely distributed in the northwest part of China. Although Chinese loess is relatively homogenous in comparison to other soils worldwide, its properties unavoidably exhibit inherent spatial variabilities as a result of complex geological and geo-environmental processes that occurred during loess formation process and various factors influencing the formation processes. Furthermore, the inherent regional heterogeneity of loess features may also account for the occurrence of numerous spatially varied geo-disasters in Shaanxi Province, China. However, few studies have examined the spatial variabilities of loess properties in a comprehensive manner, which is critical for probability-based geotechnical/geological design and analysis, risk assessment and management of geo-hazards on the Loess Plateau, etc. In this study, a comprehensive literature review and numerous laboratory tests are conducted, with a collection of 25, 82 and 177 loess profiles respectively for sandy, silty, and clayey loess to examine their properties in Shaanxi Province, China. This study evaluates and reports the typical ranges of loess properties, including property mean, coefficient of variation, most suitable probability distribution functions and spatial correlation lengths along the vertical direction. Notably, when site-specific data of loess properties are not available for the geological/geotechnical projects at hand, the typical ranges of loess properties, including index properties, strength properties and deformation properties, can be used as guidelines for approximation. Besides, the statistics and correlation length to be reported in this study can be used as prior knowledge for Bayesian characterization of loess variability. This is particularly advantageous for small to medium-sized hazard-prone regions, where the amount of data on loess properties is often insufficient to generate meaningful statistics for use in geotechnical or geological design and analysis to prevent and mitigate geo-disasters.