4 Strong 5,000--8,000 | 2.7 -- 5.9 
V. Strong | 8000--13500 | 6.0 -- 10.00 
Severe > 13,500 >"10.0 
  
2.2.2 Soil Erosion Rate: It was ranked in five 
grades which was defined by CNHDM as well, 
and shown in the Table 2. 
Table 2. Erosion Rate 
  
  
  
  
  
  
  
  
  
  
Rank | Description Soil Layer 
Loss Time (Yr.) 
1 None >1000 
2 Relative Danger | 100 -- 1000 
3 Danger 20 -- 100 
4 V. Danger <20 
S Destroyed All gone 
  
2.3 Research Strategy 
In order to output a soil erosion map at a large 
scale via remote sensing and GIS technologies, 
the following strategies were adopted: 
1) Considering map usage, several factors appear 
on a map: 
. Must be match with 1:100,000 scale; 
. Must can be obtained from remotely 
sensed data or from the other supportive 
data; 
. Can enhance applications at county level; 
. Can suggest future change. 
2) Factors were ranked not on theory, but based 
on practical experience. In the other words, the 
research was application oriented, so even if a 
known theory of soil erosion existed it was not 
adopted if it could not be implemented smoothly 
in practical mapping. Hence the ranking system 
for each factor: 
» Must be identifiable on remotely sensed 
data; or identifiable on the other 
supportive data; 
° Must be identifiable in the field sampling 
sites; 
. Be as simple as possible for easy 
modeling; but 
818 
* Establish sufficient ranks for the other 
application modification. 
3. FACTOR SELECTION AND 
GENERATION 
Factor selection is very important to soil erosion 
mapping on a large scale. In China National Soil 
Erosion Survey in the 1980's, landform was used 
as a very important factor. The authors, however, 
think the landform factor has little influence on 
large scale mapping, so this research did not adopt 
this factor. 
3.1 Rainfall Factor 
In the USLE, the rainfall factor was termed as the 
Rainfall Index which was calculated based on the 
maximum rainfall period. But as the rainfall type 
varies, the index is hard to determine. The authors 
had no sufficient experimental data and time to 
work out which time span, 15 minutes, 30 minutes 
or 60 minutes, is optimal for the index calculation, 
and which index, Iso, Ico or Iis, could satisfy the 
mapping needs in an area of high local relief. 
The rainfall factor was simplified, in this research, 
in terms of rain coefficient which were defined in 
relative values. The research area was delineated 
into five subzones according to their annual 
precipitation. Any subzone can be taken as a 
standard, so the rainfall coefficient in this standard 
subzone was set as 1.0. Then the coefficients in 
the other subzones were set according to annual 
precipitation increment against the standard 
precipitation. 
3.2 Soil and Parent Material Factor 
The soils, though complicated in types, were 
ranked into eight grades according to their 
resistability to erosion, rather than the erodability 
which was adopted by USLE. This inverse setting 
was purely for easier manipulation in modeling. 
The ranking table is illustrated in Table 3. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996 
  
  
  
  
  
  
  
  
  
  
This ra 
fixed: 
rankin, 
compli 
modeli 
made 
geologi 
ARC/T 
format 
3.3 Slo 
Slope 
ranking 
the soil 
map. Ii 
would 
fields a 
the mar 
than soi 
up slop 
ranking