International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
drainage areas and the average of the discharge of an order of 
zero ( 
800 
  
Fig. 10. Drainage map of the Madeira river basin 
Number 
  
10000 
1000 B. 
100 
10 “+ 
  
  
  
Zem Fist Second Third Fourth Fifth’ Onder 
Fig. 11. Number of branches vs. Orders 
Length fm) 
  
10000 
100 + 
  
  
  
Fist Second Thid O xier 
  
In 
22 0 mer 
Fourth 
  
  
  
Zero First Second Thád 
Fig. 13. River slope vs. Orders 
Table 1. River characteristics for each order 
  
  
  
  
  
  
  
  
  
  
Order 0 (estimate) 1 2 3 4 5 
Number of the branches 1191 227 42 8 3 1 
Length of the branches (km) 48.4 82.5 139.6 | 428.3 | 507.6 | 312.4 
River slope 1.573 1570 | 1.567 | 1.562 | 1.553 | 1.556 
  
  
Table 2. Estimated characteristics of the order zero 
and the average of discharge 
  
  
  
  
  
  
  
Length of the branches Drainage area | Average of the discharge 
(km) (km?) (m/s) 
Estimate of the order of zero 48.4 632.6 13.5 
isible river characteristics 
Kvisible fivcr enaracte 31.5 628.0 13.4 
estimated from SAR 
  
  
Length (km) 
  
100000 p——á— 
10000 
  
  
10000000 
Ana (m2) 
Dschame (m 
  
100E*13 [e E 
  
1D0E+12 
100E+11 
  
  
100E410 
10000 100000 10000000 
Ara km2) 
1000000 
Fig. 15. Discharges vs. Areas 
4. CONCLUSIONS 
We analyzed the JERS-1/SAR images about 12 scenes from 
1993 to 1997 in the Amazon, and found out the characteristics 
of the river shapes with the river geomorphology correlated the 
discharge well. Namely, as a result of the Fourier and wavelet 
analyses, the more discharge had the less spatial frequencies, 
while the less discharge had the higher spatial frequencies. In 
particular, we found that the continuous wavelet analysis was 
the best method to estimate the river discharge. Moreover, by 
the river geomorphology or Hack’s and Horton's rules with the 
macro scaled drainage map, we could estimate the river 
characteristics of the order of zero, and they almost 
corresponded with the invisible river characteristics estimated 
from SAR images. However, for the reason the characteristics 
were averaged for each order, the precision had some of error. 
Therefore, we could obtain the difference between some rivers, 
but a very small change such as a seasonal change was hard to 
be derived. In the future, we will analyse and estimate the 
seasonal change of its discharge with the precipitation data and 
brightness of the river. Accordingly, by the analysis of the river 
characteristics, we could estimate the discharge of the narrow 
branches in the Amazon basin, and monitor the water budget 
and soil runoff. 
References 
Takako Sakurai-Amano, Joji lisaka, and Mikio Takagi, 2000. 
Detection of narrow open-water channels from JERS-1 SAR 
images of Amazon forests, Proc. of SPIE's Second International 
Asia-Pacific Symposium on Remote Sensing of the Atmosphere, 
Environment, and Space, Sendai, Japan, pp.120-130. 
Takako Sakurai-Amano, ef al, 2001. A year change and 
monthly division precipitation of a river in the Amazon forest, 
Soc. of Photogrammetry and Remote Sensing, Toyama, Japan. 
Shigemi Takagi, 1974. River Morphology, Kyoritsu publication. 
892 
fo 
Fi 
the 
CO! 
an 
Ce 
Or 
roc 
cre 
Cal 
int 
rin 
to 
CO! 
Sul 
alr 
the 
Re 
CO! 
po