Optimization of permeability estimation in the Loess Plateau: Insight into the quantification of structural disorder

Abstract

Reasonable quantification of soil structural characteristics has the potential to significantly improve the understanding of soil hydrological processes. Herein, this paper examines strata seepage mechanisms under soil structural factors, and introduces the configuration entropy to solve the uncertainty of permeability prediction caused by the structural disorder. The results show that the loess layer has a higher hydraulic conductivity than the adjacent paleosol layer, and the differences in the permeability between the loess and paleosol layers decrease with depth. The configuration entropy, as a structure probability function, is a quantitative index reflecting the characteristics of porous media. The configuration entropy of the strata exhibits significant periodic characteristics, and the configuration entropy of the loess is greater than that of the paleosol. This indicates that the disorder of the loess layer is higher, water passages are more abundant, and and the structure is more unstable. The paleosol layer has a relatively uniform particle distribution, lower permeability, and stable structure. This difference in soil structure is caused by the influence of climate and sedimentary environment. Since the configuration entropy can well reflect the characteristics of anisotropy and structural disorder, the configuration entropy is used to improved the Kozeny–Carman (KC) formula. A comparison between the measured and fitted values show that the improved prediction accuracy of the permeability is higher than that of the conventional KC formula. A new auxiliary index for permeability prediction and an improved KC formula are proposed in this study, which is excellent for quantifying the structural complexity parameters and predicting the hydraulic conductivity.