Effect of transport layer system on suction distribution for tropical residual soil slopes

Abstract

Substantial surficial deposits of many tropical climate regions are covered by tropical residual soils. The weathering process forms a layered sloping soil of Grade VI and Grade V soil mantle with variable hydraulic conductivities, ksat which creates capillary barrier effect at the interface of Grade VI and Grade V soil layers. Although the capillary barrier effect impedes downward water infiltration, the water diversion capacity is limited and the moisture content increases at the interface that can lead to potential slope failure. Hence, this study investigates the effects of employing transport layer to increase the diversion capacity at the interface of Grade VI and Grade V soil layers via laboratory experiments, field monitoring and numerical modelling. A laboratory physical slope model was developed to perform infiltration tests with five (5) configuration schemes of Grade V and Grade VI soils sandwiched by four (4) different types of transport layer, i.e. Gravel, Drainage Cell (DC)+Gravel, Sand and DC+Sand, and without transport layer. A total of thirty two (32) infiltration tests were performed in this study. Three research plots i.e. a control plot without transport layer and two (2) plots with sand transport layer as well as with gravel transport layer were constructed and instrumented to monitor rainfall, runoff, amount of diverted water and matric suction distribution. The monitoring was performed during wet period from September 2014 to January 2015 where the soil experienced high water content but low matric suction. Subsequently, the two-layered slope with and without the transport layers was numerically simulated using a finite element method to validate the field data and to determine the best modelling scheme to represent the residual soil slope model with transport layers. The results of the laboratory experiments clearly shows that the transport layer sandwiched between the Grade VI and Grade V soil layers was capable of diverting the infiltrating water above the interface. There was a significant increase of matric suction measured at the interface of soil layers with DC + Gravel transport layer as compared to that without the transport layer especially for 2-hr and 24-hr rainfall intensities while the effect was insignificant for 7-day rainfall intensity. Field monitoring also indicates that the initial matric suction value in the control plot responded to the infiltration and reached a matric suction value that corresponds to the breakthrough suction of 5.0 kPa after series of rainfalls. However, the initial matric suction values were relatively maintained in the plots with transport layers to indicate that the transport layer played the role of increasing the amount of diverted water at the interface. The finding was supported by the results of amount of diverted water collected at the research plots. Continuum model is capable of modelling the effects of employing transport layer at the interface by subdividing the layer into multiple isolated zones with different average ksat. The results of the analysis demonstrated that the capability of transport layer to maintain the matric suction and to divert water was governed by the contrast in the ksat where the higher ksat of gravel shows a better performance as compared to sand. A combination of initial suction at 30 kPa, 0.5 m thickness of sandy silt (Grade VI), 0.3 m thickness of gravel transport layer and 21° slope angle resulted in a diversion length of more than 15 m. However, the lower initial suction value due to rainfall infiltration during wet period yielded a shorter diversion length.