Soil Suction Dynamics in Vegetated Soil: From Deterministic to Probabilistic Analysis with Electrical Resistivity

The importance of soil suction measurement in site characterization is very well recognized. In the field, soil suction is typically measured using tensiometers, which give point information. In vegetated soil, the change in soil suction is difficult to predict in a deterministic manner using point-based information due to spatial variability and heterogeneity. Large-scale field soil suction characterization is even more challenging. On the other hand, electrical resistivity is an important parameter in geophysical engineering, which is acquired by tomography surveys in the field. Therefore, probabilistically correlating soil suction with resistivity in the field might render a representative site characterization. The major objective of this paper was to correlate soil suction to electrical resistivity based on field investigation results using both deterministic and probabilistic methods. To pursue the objective, tensiometers were installed at multiple depths in multiple vegetated soils, and resistivity tests were conducted at the same location periodically. An analytical relationship between the field-measured suction and resistivity was derived, and the degree of strength of the correlation between suction and resistivity was measured using the non-parametric statistical measure: Spearman Rank Correlation Coefficient. The correlation was found to be highly dispersed at a resistivity value of more than 60 Ω-m. The higher resistivity values were congregated into small clusters. The Gaussian distribution theorem was then used to correlate the soil suction at higher resistivity values probabilistically. The results indicated that the probabilistic method better portrays the relationship between soil suction and resistivity in vegetated soil. Our findings provided the high potential of the electrical resistivity method in contributing to more effective monitoring and estimating soil suction in vegetated soil.