Capturing the hole effect in Qiantang River alluvial silt by cone penetration test

Soil spatial variability exhibits significant regional characteristics arising from various geological processes, weathering effects, and stress history during sedimentation. The spatial variability can influence the performance of geotechnical systems, and accurately capturing it is essential for effective reliability analysis. This study first compares the performance of the auto-correlation function and variance reduction function under different fitting strategies. Subsequently, the sampling strategies that yield the best performance in estimating the scale of fluctuation (SOF) are identified. After that, cone penetration test data are collected from sites along the Qiantang River to quantify the spatial variability parameters of the alluvial silt. Analysis reveals that the SOF and its associated uncertainty generally increase with expanding investigation domain length. The mean SOF rises from 0.691 m to 1.225 m across scales from 5 m to 15 m. The hole effect describes a spatial autocorrelation function that exhibits a non-monotonic, cyclical pattern as lag distance increases. A significant hole effect, identified in 75.6% of cases with SOF values ranging from 0.190 m to 1.135 m, may be associated with rhythmic tidal sedimentation processes. The cosine Whittle-Matérn model successfully captures the finer features of the spatial correlation structure, including the SOF, smoothness, and hole effect.