geotechnical properties, and strength parameters of the soil on slope instability is
described in the time-invariant spatial distribution map.
3. The process-based distributed hydrological model which is described in the next
section is calculated using the observed AmeDAS rainfall data and Geospatial
data from step 1.
4. Evaluation of the hydrological model performance through calibration of the
hydrological model response was carried out using observed stream flow
discharge.
5. The slope stability evaluation was analysed by comparing the spatial pattern
of landslides inventory data with the pattern of simulated time-invariant slope
stability classes.
6. Based on the evaluated time-invariant slope stability distribution map, the longterm spatial dynamic and time-varying of potential slope instability map can be
drawn. The hydrological model predicts the dynamic of soil saturation in each
grid element, which is then used to update the state of relative soil saturation and
to assess local slope instability for whole areas defined as potentially stable/
unstable.
7. Shallow landslide probability of occurrence is simulated from the Spatial
Dynamic and Time-varying of Potential Slope Instability Map.
8. Finally, the shallow landslides hazard map can be created with the simulation of
shallow landslide occurrences probability.
11.2.2 Physical Based Hydrological Model
A physically based hydrological model coupled with a slope stability model was
developed to assess shallow landslide risk analysis over large areas. The distributed
hydrological model is the grid-Cell Distributed Rainfall Runoff Model Version
3 (CDRMV3) which was developed at the Innovative Disaster Prevention Technology and Policy Research Laboratory, DPRI, Kyoto University. The CDRMV3
model solves the Kinematic wave equation using the Lax-Wendroff scheme at
every node of each cell (Kojima et al. 2003). An automatic calibration program
using the Monte Carlo method was added to the evaluation of model performance
and uncertainty analysis of the CDRMV3 (Sayama et al. 2003; Apip et al. 2010).
Using a steady state assumption, a lumped sediment-runoff model was developed
and applied by Apip et al. (2012) based on the CDRMV3 model structure.
The catchment topography is taken from the digital elevation model (DEM)
which is divided into square grid-cells. A square area with four node points is called
a grid-cell. The analyzed catchment is calculated as a network of grid-cells. The
flow of each grid-cell receives the flows from upper grid-cells and direct rainfall.
Connected grid-cells receive flow based on the drainage path defined by selecting
the steepest direction from eight-directions. Discharge and water depth flow to the
next grid-cell according to the predefined eight-directional flow map and routine
order determined in accordance with DEM and river channel network data. Flow is
226
P. Luo et al.
described in the time-invariant spatial distribution map.
3. The process-based distributed hydrological model which is described in the next
section is calculated using the observed AmeDAS rainfall data and Geospatial
data from step 1.
4. Evaluation of the hydrological model performance through calibration of the
hydrological model response was carried out using observed stream flow
discharge.
5. The slope stability evaluation was analysed by comparing the spatial pattern
of landslides inventory data with the pattern of simulated time-invariant slope
stability classes.
6. Based on the evaluated time-invariant slope stability distribution map, the longterm spatial dynamic and time-varying of potential slope instability map can be
drawn. The hydrological model predicts the dynamic of soil saturation in each
grid element, which is then used to update the state of relative soil saturation and
to assess local slope instability for whole areas defined as potentially stable/
unstable.
7. Shallow landslide probability of occurrence is simulated from the Spatial
Dynamic and Time-varying of Potential Slope Instability Map.
8. Finally, the shallow landslides hazard map can be created with the simulation of
shallow landslide occurrences probability.
11.2.2 Physical Based Hydrological Model
A physically based hydrological model coupled with a slope stability model was
developed to assess shallow landslide risk analysis over large areas. The distributed
hydrological model is the grid-Cell Distributed Rainfall Runoff Model Version
3 (CDRMV3) which was developed at the Innovative Disaster Prevention Technology and Policy Research Laboratory, DPRI, Kyoto University. The CDRMV3
model solves the Kinematic wave equation using the Lax-Wendroff scheme at
every node of each cell (Kojima et al. 2003). An automatic calibration program
using the Monte Carlo method was added to the evaluation of model performance
and uncertainty analysis of the CDRMV3 (Sayama et al. 2003; Apip et al. 2010).
Using a steady state assumption, a lumped sediment-runoff model was developed
and applied by Apip et al. (2012) based on the CDRMV3 model structure.
The catchment topography is taken from the digital elevation model (DEM)
which is divided into square grid-cells. A square area with four node points is called
a grid-cell. The analyzed catchment is calculated as a network of grid-cells. The
flow of each grid-cell receives the flows from upper grid-cells and direct rainfall.
Connected grid-cells receive flow based on the drainage path defined by selecting
the steepest direction from eight-directions. Discharge and water depth flow to the
next grid-cell according to the predefined eight-directional flow map and routine
order determined in accordance with DEM and river channel network data. Flow is
226
P. Luo et al.