Experimental study on the water retention behavior of intact loess under mechanical wetting and hydraulic wetting

Natural loess usually experiences mechanical wetting and hydraulic wetting in situ conditions. In this paper, the difference of soil water retention curves of natural loess along mechanical wetting and hydraulic wetting was examined and discussed. The Gallipoli's, Tarantino's, and Hu's models were adapted to fitting a three-dimensional surface of the degree of saturation S-r against suction s and void ratio e to evaluate the performance of those models when describing the water retention behaviors of intact loess. The SWRCs presented were obtained by using a Fredlund SWCC device for hydraulic wetting and a suction-controlled oedometer for mechanical wetting. The results show that the degree of saturation increases with the decrease in the void ratio under the constant suction condition. Furthermore, the experimental data of mechanical wetting and hydraulic wetting in (e, s, S-r) space can be reasonably fitted by the Gallipoli's, Tarantino's, and Hu's models; however, the two fitted surfaces of the mechanical and hydraulic path do not coincide with each other. The result indicates that the SWRC is not only affected by the void ratio but also related to the way of wetting. Moreover, those models can accurately predict the variation of the saturation under suction-controlled oedometer tests. The results in this study would contribute to clarifying the difference in SWRC of mechanical wetting and hydraulic wetting of intact loess and to broadening their application of those models in the loess region.

Automated summary

The difference in soil water retention curves of natural loess under mechanical wetting and hydraulic wetting was examined and discussed. Three models were used to evaluate the performance of describing the water retention behaviors of intact loess. The results showed that the saturation increased with the decrease in void ratio, and the fitted surfaces of the mechanical and hydraulic path did not coincide.