Hydraulic conductivity over a wide suction range of loess with different dry densities

Abstract

Experimental research into the hydraulic conductivity curve (HCC) of unsaturated soil is limited due to the inherent challenge associated with labor, cost, and time. Typically, the HCC is estimated using the soil water characteristic curve (SWCC) based models and saturated hydraulic conductivity (SHC). However, the efficiency of the SWCC-based model is rarely assessed, and the influence of soil density and pore structure on HCC remains incomplete due to limited experimental data. To address this gap, this study employs an innovative filter-paper-based column method, which can measure the HCC over a wide suction range (e.g. 0−105 kPa), to capture the HCCs of both intact and compacted specimens with varying dry densities. The efficiency of two typical SWCC-based models is assessed using the measured data. Meanwhile, the mercury intrusion porosity (MIP) technique is employed to obtain the pore characteristic (i.e. pore size distribution (PSD)) and a method of predicting the HCC using the PSD data is proposed, emphasizing the dominant role of the pore structure in shaping the HCC. The results reveal that the dry density's influence on the HCC is primarily observed within the low suction range, corresponding to variations in the dominant and large pores. In the high suction range, the HCCs align along a linear trajectory when plotted in a log-log format. A notable finding is the overestimation of the HCC obtained from the SWCC-based models using the measured SHC. When the SHC is regarded as a fitting parameter, good agreement is achieved. The adjusted SHC value is typically 0-1 order of magnitude lower than the measured value, and this discrepancy diminishes as dry density increases. On the other hand, the proposed PSD-based model performs well with the measured SHC data. Caution is exercised when using the SHC to estimate the HCC for modeling water movement in partially saturated soil.