Characteristics Of Loess Saturated Hydraulic Conductivity In Landslide Of The South Jingyang Plateau

In this paper, we focused on loess landslides of the south Jingyang Plateau in Shaanxi Province. Under the field investigation, we excavated five trenchs in different landslides and two adits in the same landslide. Soil samples were collected. The horizontal and vertical saturated hydraulic conductivity of these loess samples can be got from the variable head permeability test. Negative exponential model was applied to study the saturated hydraulic conductivity changes with the depth. Traditional statistics and geostatistics were applied to research the spatial variability of loess saturated hydraulic conductivity, providing a scientific basis for the influence of slope stability. The main conclusions are as follows:(1)The saturated hydraulic conductivity of trenchs is in the range of 10-2m/d to 101m/d, while the saturated hydraulic conductivity of adits is less than trench’s, range from 10-4m/d to 101m/d. The loess saturated hydraulic conductivity has obvious anisotropy: vertical hydraulic conductivity is bigger than horizontal hydraulic conductivity in trenchs, but the results are opposite for that of adits.(2)The loess saturated hydraulic conductivity decays with depth as a whole. Malan loess’ s saturated hydraulic conductivity is greater than Lishi loess’ s. Upper saturated hydraulic conductivity is bigger than lower’s in Lishi loess. Negative exponential models fitting the variation of hydraulic conductivity with depth are given for different strata of five trenchs. The loess saturated hydraulic conductivity of Malan loess increases with depth, while that of Lishi loess reduces with depth, but there are specific areas.(3) Based on the method of traditional statistics, the loess saturated hydraulic conductivity is in a strong degree of variation. Discrete degree of horizontal hydraulic conductivity is greater than the vertical hydraulic conductivity, and they are all in lognormal distribution.(4) Geostatistics analysis results show that Gaussian model is more suitable for fitting loess saturated hydraulic conductivity in trenchs, and Exponential model is more suitable for fitting loess saturated hydraulic conductivity in adits. The loess saturated hydraulic conductivity has intermediate spatial autocorrelation: spatial autocorrelation loess vertical hydraulic conductivity are greater than those of horizontal hydraulic conductivity.