le 
2 
v. 
of. 
of 
—— 
  
  
va «4 
^ ADS A AU M 
- Ud J 
ARMS. v^ CK RCH MML.SM | 
T 
     
Topological 
   
   
   
  
Map i> 
| DI: --D 
Height point | Calculate 
Contour line : Aspect and|i 
|l: Gradient 
  
  
  
uli + 
Prats ea] |G | | 
rr 
> ri | | 
  
Geological 
Map 
Character 
Landslide 
desribution 
Map 
| t IE 
Landslide 
de a or LS construction of database 
   
  
  
| Overlay 
  
  
  
  
  
  
  
    
  
  
  
      
  
  
  
  
  
  
  
E q Topological 
— x [Sm Map 
T E nn 
R Landslide 
A D desribution 
FOE : Map 
>> ST 
T Overlay 
| + 
O 
> N 
O 
F 
> Y 
F Renew 
A landslide 
C destrinution 
T Asupport 
O Attempt to 
R landslide 
7 points 
  
  
  
Fig.2 Flow chart of this study 
2. Study method 
This study is carried out following the 
procedures shown in Fig. 2. 
2.1 Construction of database 
  
Since there is no digitized geographical 
information available for the study area, we 
constructed the database from the topographic map 
and geological map using a digitizer. The digitized 
geographical information is as follows: 
2.1.1 Landslide distribution: As for the 
distribution of landslides occurred already, we use 
the landslide distribution map at a scale 1:50000 
(The Public Works Research Institution, 1969) made 
in 1969. . As for the landslide distribution 
information, the shape of landslide is digitaized 
and converted into polygon data. 
  
2.1.2. Digital terrain model (DIM): DIM 18 
constructed by digitizing contour lines and 
altitude points from topographic map at a scale 
1:25000. And then DTM is converted into TIN 
  
(Triangulated Irregular Network). Topographical 
information is made from TIN, and gradient and 
aspect are calculated therefrom. The TIN is 
converted into polygon data, and the topographical 
information is added to TIN polygon data. 
2.1.3 Land information: Landslide exerts influences 
on the various phenomena on land. Especially in 
many cases, ponds and paddy fields are located on 
the gentle slope land of landslide area. In case of 
this study area too, there are many paddy fields 
and ponds on the area where landslide already 
occurred. In this study, therefore, land 
information of the shape of paddy field and pond is 
digitized from topographic map at a scale 1:25000 
and is converted into polygon data. 
2.1.4 Lithologic. character: In. many cases, the 
geological age is used as an indicator and 
stratigraphic classification is indicated on the 
geological map. However, in considering the 
relation between landslide and geology, we place 
our focus not on the stratigraphic classification 
  
651 
but on what kinds of 1lithologic character the 
stratigraphic classification is composed of. Also, 
there are some cases where different stratigraphic 
classification may have the same lithologic 
character. Therefore, we digitize the stratigraphic 
classification to the database from geological map 
at a scale 1:50000. After the digitized data is 
converted into polygon data, main lithologic 
character of each stratigraphic classification is 
added to the data as attribute information. 
2.2 Extraction of factor information 
  
Each of geographic information compiled as a 
database following the procedures as stated above 
is overlaid with the distribution information of 
landslides occurred already. Overlay is carried out 
in an aim to study the relation between landslide 
and each kind of geographical information. As a 
method to study the relationship, we compare the 
area ratios which are occupied by each category 
contained in the geographical information within 
the study area with those area ratios within the 
landslide area. As a result of comparison, the 
geographical information which recognized 
difference in the ratio between the landslide area 
and the total study area is used for the analysis 
as an effective factor. 
We used to use much time and labor to 
quantitatively analyze the factor information of 
landslide in the past, but it is now possible to do 
this analysis far more effectively by using GIS. 
2.3 Attempt to extract landslide points 
  
The factor information to be used for analysis 
is created in an output layer by overlay. As a 
result, the information organized in one layer is 
added with the score obtained by the area ratio of 
each category of factor information within the 
landslide area. Finally, we attempt to extract 
landslide area by picking up the areas which have 
higher scores. 
2.4 Attempt to renew the landslide distribution 
  
information 
The information extracted through the above