by 
nd 
ng 
er 
dy 
he 
of 
in 
re 
ds 
of 
3) 
dy 
  
Uv 
e 
o 
20 
m 
nf 
AVERAGE:77 
  
h90i 4» 01O - 0000 C — 
OOOOOOOOOO 
  
  
0 400 800 (ha) 
T 0O T A 1 A R E. A 
(1) Score destribution in londslide area 
Fig.3 Result. of getting 
  
  
  
SCORE 
110 
100 
90 
80 1305 
70 AVERAGE:70 
60 
50 
40 
30 
20 
1000 2000 (ha) 
T. 0 T A L A R E A 
(2) Score destribution in total study areo 
score of overlaid polygon. 
  
  
ut 
0 
EXTRACTED AREA 
100% 
80% 
60% 
40% 
20% 
0% 
  
  
Fig.5 Rote of extracted 
  
1 
F R E O0 U E N 
0 
150 
0 
CY 
area in each landslides. 
  
3.2 Attempt to extract landslide areas 
  
The factor information used for the analysis of 
3.1 was organized into one :layer ‘by polygon 
overlay. The polygon created as a result of polygon 
overlay was added with the score worked out for 
each factor from its relationship with the 
landslide, The extraction of landslide areas is 
carried out basing on these score. The extracted 
results are discussed from the viewpoint of 
coordination with the already-occurred landslide 
areas. 
3.2.1 Establishment of threshold: It is necessary 
to give a score for each category and add weight to 
the factor information organized into one layer. At 
present, however, there is no effective method to 
do so which is both objective and quantitative. In 
this study, therefore, we worked out the ratio of 
area occupied by each category within the landslide 
area and discussed thereon the relationship between 
the factor and landslide. Namely, the difference of 
area ratio within the landslide area can be 
considered as the difference of weight among the 
categories. For example, lithologic character at 
the total area of landslide Shows that 
approximately 11% of "colluvial and debris" and 
approximately 2% of. "gravel, sand and silt" are 
distributed on the already occurred landslide 
areas, Within the total study area, on the other 
hand, the share of "colluvial and debris" is 
approximately 5% and that of "gravel, sand and 
silt" is approximately 6%. Thus, if the area ratio 
of categories related to the landslide, the area 
ratio changed between the landslide area and the 
total study area. Therefore, the value of percent 
of each category at the landslide area can be used 
as the weight. This way of weighting is considered 
to be more coinciding with the actual state. 
Also, it is impossible to ignore the relation 
between landslide and ponds/paddy fields. However, 
the database of ponds and paddy fields was not the 
one created as the result of classification made to 
  
653 
the total study area, but it is a result of 
extraction the land information. Therefore, as the 
method to add weight to the ponds and paddy fields, 
we decide to calculate. the ratio of area occupied 
by each factor within the landslide area against 
the area of each factor within the total study 
area, and to add the percent value as a score. 
Also, these scores are used as the threshold 
value in extracting the landslide areas. The 
information to which the weight addition was 
completed is overlaid with the landslide 
distribution information. Fig. 3 shows the 
distribution of scores within the total study area 
and at the landslide areas. The average score in 
the already occurred landslide areas is 77 points, 
while the average score of total study area was 70. 
Now, the difference of score distribution between 
the total study area and landslide area is that the 
score distribution at the landslide area is lower 
than that in the total study area below the 
average. Namely, the landslide areas are 
distributing at the high score areas within the 
total study" area. Basing on this result, we 
extracted the polygons which have the score of 77 
or higher, which is shown in Fig. 4. 
3.2.2 Analysis result: Here, we will evaluate the 
information extracted through the above-mentioned 
process. Fig. 5 shows the ratio of region with the 
score 77 or higher against each landslide area. 
There are 672 landslide points in the study area. 
Out of which, the landslide points which are 
located within the extracted area are 599 
(including the point only a part of which is 
included in the extracted area). Namely, the 
landslide points in the extracted area is 88% of 
all the landslide points in the total study area. 
Also, the landslide points 60% or more of which is 
occupied by the score 77 or higher are 366. Namely, 
the landslide points 60% or more of which is 
occupied by the score 77 or higher is 54% of all 
the landslide points within the total study area.