ate) 
ate) 
isopleth map is described by Wright(1974), 
DeGRAFF(1988) and others. 
The first step of the map preparation is an 
inventory of exin landslides by aerial 
photographs and field  verification(Fig.1). A 
transparent orthogonal grid sheet with 1.3 or 
1.6 cm intersection spacing is placed over the 
landslide inventory map(Fig.2-A). A counting 
circle(2.5cm diameter) enclosing an inscribed 20 
x20 per 2.5cm grid, is centered at each 
intersection. The number of points within the 
circle over landslide is divided by total number 
of grid points within the circle and multiplied 
by 100 to yield a percentage value. This 
process is repeated until that all intersections 
have percentage values(Fig.2-B). Contour lines 
(isopleths) are then drawn through numbers to 
represent in given interval or values(Fig.2-C). 
The isopleth map was used in two ways. 
Firstly, it serves as initial means for landslide 
hazard assessment. Secondly, it enable to 
recognize some  landslide-susceptible slopes by 
overlay technique of isopleth map and slope 
map in GIS system. In the study area, 
landslides occur mainly on slopes of less than 
30 degrees(Fig.3-B). In the other regions(more 
than 40 degrees), they are not taking place 
frequently, but the risk is high. 
DeGRAFF(1985) has presented landslide 
susceptibility categories identified on  isopleth 
map. The overlay results of two  maps(fig.3) 
permits us to determine new categiries in the 
study area(Table 1). These categories are used 
to develop a stability rating system as being 
mentioned in the next chapter. Large scale 
mass movement is not recognized on the 
isopleth map, but only small ones are seen 
in hilly or mountainous terrain. 
ANALYSIS OF GEOLOGIC HAZARDS 
(LANDSLIDES) USING GIS 
Slope failures depend on four kinds of 
following variables(Hunt,1983, Walker et al, 
1987): topography(slope inclination and height), 
eology(material structure and strength), seepage 
orces and runoff quantity, seismic activity. On 
the basis of these elements, nine factors(Table 
2) were examined to establish a stability rating 
system in accordance with environmental 
characteristics of this study area. The ratings 
were developed through the evaluation of each 
range for each factor with respect to the other 
factors. They range from 1 to 7. These 
suggested ranges may be adjusted by other 
users when varying conditions are indicated by 
the data(Griner, 1989). Seventeen data layers 
(Table 3) were prepared by scanning, 
digitizing, vectorizing and editing processes to 
analyse these factors from various source data. 
Slope and Landslide Frequency 
Considering the relationship between slope and 
landslide occurrences, slope angles should be 
divided into several classes. The analyses by 
overlay technique(slope and isopleth maps) have 
already been illustrated in Table 1 and Fig.3. 
By this method, the reasonable ranges of Sone 
and landslide frequency could be set up. The 
ratings of landslide frequency are presented in 
Table 1, and those of slope are as the 
677 
  
Tel 7 pA 
e 
Sel Gel si 
p 
^ - 
da Cr 
go ge ? ; 
4 
‘ 
” 
^ 
(c) Case3 (R1) 
^ 
  
(b) Casze2 (R2) 
(a) Casel (R3) 
  
  
(Tel: topographic contour, Gel: groundwater 
table, Sel: basal surface line of soil, Wel: top 
surface line of bedrock) 
Fig.4 Analysis of groundwater table 
followings; 1(0°-10°), 2(10°-20°, 3(20°-30°, 4(30 
°-40°), 5(40°-50°), 6(above 50°. The slopes of 
higher than 60 degrees and landslide 
frequencies of more than 35% are not almost 
present. 
Groundwater Level 
The rise of groundwater level severely increases 
seepage forces, which may reduce the resisting 
forces along the failure surface or increase the 
driving forces(Hunt, 1983). The groundwater 
level is affected by rainfall accumulation 
(infiltration), reservoir filling and other factors 
(influences of rainfall are examined in the next 
chapter). For this work, the locations of 
groundwater table were classified into three 
cases according to their  levels(Fig.4). For 
example, when the level is located in soil 
zone, the rating is high(3). A data layer(Fig.5) 
was created by processing geologic information 
of subsurface such as soil depth, weathering 
zone, bed rock and groundwater level, using 
ARC/INFO, IDRISI and our programs of 
pascal language. This information was prepared 
from 271 test boring data(Soil investigation 
report, 1990, Geological report for electrical 
railway, 1989). 
Vegetation 
Roots reinforce the soil, increasing soil shear 
strength. They also bind soil particles at the 
ground surface, reducing their susceptibility to 
erosion. Foliage intercepts rainfall, causing 
absorptive and evaporative losses that reduces 
rainfall available for infiltration(Walker et al, 
1987). From these viewpoints, vegetations in 
the study area were divided into four groups 
according to their density; heavy vegetation(1), 
light vegetation(2), grass or very light 
vegetation(3), no vegetation(4). 
In order to analyse vegetations, land cover/use 
maps were prepared by PCA (Principal 
Component Analysis) technique and by 
clustering methods with Thematic Mapper 
images(Oct. 1987, Sep.1988). Interpretation of 
aerial photographs supported the shadow parts 
of TM images, streams and roads. The TM 
images was analysed using a classification 
system of U.S.Geologocal Survey(Anderson et 
al, 1976). 
Light or no vegetations and barren lands are 
hardly distinguishable from agricultual lands, 
which are mostly located on slopes of less 
than 10 degrees in the study area. Therefore, 
the slopes of less than 10 degrees are excluded 
from the overlay result of slope and land 
coverfuse maps. Rest of the area is assumed