ey 
ich 
on 
  
  
Table 3. Distribution of hazardeous zone in different elevation (ha). 
  
  
  
  
  
  
  
No Hazard Rank 1&2 Rank 3&4 Total 
0- 100m 909.54 0.00 0.00 909.54 
-200m 36.81 0.00 0.00 36.81 
- 300m 36.72 0.00 0.00 36.72 
- 400m 39.60 0.00 0.00 39.60 
- 500m 34017.57 267.30 689.67 34974.54 
- 600m 31149.27 667.17 1057.95 32874.39 
- 700m 17922.06 49.68 202.77 18174.51 
- 800m 7417.98 18.54 0.09 7436.61 
- 900m 1577.88 0.00 0.00 1577.88 
-1000m 147.24 0.00 0.00 147.24 
Total 93254.67 1002.69 1950.48 96207.84 
Table 4. Distribution of hazardeous zone in different slope (ha). 
No Hazard Rank 1&2 Rank 3&4 Total 
- 5 deg. 41517.18 241.56 514.17 42272.91 
-10 deg. 35507.07 633.87 882.45 37023.39 
-15 deg. 8945.64 104.85 361.26 9411.75 
-20 deg. 4666.50 18.81 116.37 4801.68 
-25 deg. 1269.27 1.62 50.22 1321-11 
25 deg.- 1349.01 1.98 26.01 1377.00 
Total 93254.67 1002.69 1950.48 96207.84 
  
Table 5. Distribution of hazardeous zone in different direction (ha). 
  
  
  
No Hazard Rank 1&2 Rank 3&4 Total 
Level 957.06 0.00 4.32 961.38 
N-NE 12925.89 391.86 1094.40 14412.15 
NE- E 5647.86 176.49 447.39 6271.74 
E-SE 12249.45 373.77 389.07 13012.29 
SE-S 20330.37 60.57 15.30 20406.24 
S-SW 18120.60 0.00 0.00 18120.60 
SW- W 5603.49 0.00 0.00 5603.49 
W-NW 6248.07 0.00 0.00 6248.07 
NW- N 11171.88 0.00 0.00 11171.88 
Total 93254.67 1002.69 1950.48 96207.84 
  
Table 6. Distribution of hazardeous zone in different vegetation (ha). 
  
  
  
  
  
  
  
No Hazard Rank 1&2 Rank 3&4 Total 
Annual 761.13 0.00 0.00 761:13 
Barley 10148.94 10.80 4.77 10164.51 
D.O.W. 9312.21 140.40 22.05 9474.66 
D.S.D1 45539.28 585.54 1149.84 47274.66 
D.S.D2 2647.08 0.00 0.00 2647.08 
Wheat 1156.41 0.00 0.00 1156.41 
O.Wood. 2922.48 5.94 4.50 2932.92 
S.S1 10553.31 106.92 222.21 10882.44 
S. S2 5388.66 140.94 411.57 5941.17 
S.S3 1880.82 12.15 135.54 2028.51 
Total 93254.67 1002.69 1950.48 96207.84 
Table 7. Distribution of hazardeous zone in different soil (ha). 
No Hazard Rank 1&2 Rank 3&4 Total 
Gips. 3454.20 371.97 899.82 4725.99 
Liti. 3635.91 630.72 1050.66 5317.29 
Rock 3242.43 0.00 0.00 3242.43 
Xero.deep 37716.03 0.00 0.00 37716.03 
Xero.slop 22187.16 0.00 0.00 22187.16 
Xerot. 20096.01 0.00 0.00 20096.01 
Total 93254.67 1002.69 1950.48 96207.84 
  
CONCLUDING REMARKS 
Using NN, the factors in the input layer and the results or 
phenomena in the output layer are conveniently 
connected through t he learning s upervisor. Comparing to 
other multi-variate analysis methods, the accuracy to the 
supervisor would be excellent, as both quantitative and 
categorical data are available with s uch standardization. 
This is a significant advantage in constructing GIS 
models. Furthermore, the network to be used as an 
evaluation model is constructed semi-automatically, and 
itis possible to apply parameters to GIS map calculations 
to draft evaluation maps. 
Using this method, however, the priority of factors is not 
shown directly. Because networks show the relationship 
between factors and phenomena as a whole. Accordingly, 
it is not available for the analysis to compare individual 
factors. 
Land evaluations are one of bases for regional planning. 
To evaluate land objectively, logical models are 
necessary. Logical models are constructed by data 
integration and analysis, and since NN is a flexible 
system for both data and applications, it would be useful 
method to construct GIS models. 
REFERENCES 
Yamamoto, Y. (et al.), 1995. The application of the Neural 
Networks to GIS in the Construction of al Land Evaluation 
Model : Land Evaluation for Grassland Development in 
Tochigi Prefecture. J. of Japanese Agriculture System 
Science, 11(1), pp.14-25. 
*Japanese with English Summary 
Tsuiki, M.(et al.), 1993. Multilayer Feedforward Neural 
Networks Construction Program NEUROS2. 
Bull.Computing center for Research in Agriculture, 
Forestry and Fishery, B(11), pp.1-33. 
*Japanese with English Summary 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996