Numerical modeling on tertiary creep behavior of extreme rainfall-induced landslides with TRMM application for landslide early warning

In help issue warning of extreme rainfall-induced landslide in tropical soils of Southeast Asian countries, it requires the study of landslide mechanism induced by generated excess pore water pressure at the sliding surface due to groundwater table rise under rainfall storm, and examination of empirical relationship between rainfall characteristics and past landslide occurrence (precipitation analysis). To investigate the tertiary creep behavior in soils found by Fukuzono, 1985 (d2x/dt2=A(dx/dt)α, where x is surface displacement, t is time, and A and α are constant), a series of pore-pressure-controlled tests on saturated sands were undertaken in the ring shear apparatus. The tests were conducted under combined condition of predefined normal stress and shear stress with pore water pressure changes to simulate the potential sliding surface condition in heavy rainfall. Sand, its mixture with clay material, and soil samples taken at actual landslide sliding surface were used for specimen. Repeated shear test for a specimen was also additionally conducted to produce reactivated motion landslides. Numerical model simulating the Tertiary creep behavior (or progressive failure) is constructed to develop a most appropriate method for landslide early warning combined with TRMM satellite rainfall data. TRMM data were selected to apply to the Japanese Soil Water Index (SWI) in distributing threshold of highly nonlinear rainfall patterns for estimating the landslide occurrence in developing regions: Southeast Asian countries, where very limited number of rain gauges is available, and there is no implemented methodology for issuing effective warming of landslides yet. It is through the plot of total water of 3 serial tank models and daily precipitation with case example of landslide disasters took place in Beichuan city, (located on the 2008 Chinese Wenchuan earthquake fault) and Hofu city, Japan which were hit by heavy rainfall attacked in 2009. Consequently, it is successful to reproduce the tertiary creep to failure in ring shear apparatus by pore-pressure-controlled test, in which obvious relationship of A and alpha values was obtained. Alpha value range is consistent with those found in previous experiments with greater A value. Additionally repeated shear tests for a specimen to reproduce reactivated landslide motion indicated less variation of alpha value, converging into alpha value of 2. Physical model to reproduce the tertiary creep is developed. Application of TRMM satellite rainfall data to soil water index was examined. And, combination with the physical model to develop early warning in Southeast Asian is under development.