42 
Landsat TM 
In situ data 
n 
Y 
f 090 
X=k’ S S(A)*R(A)*x(A) d A 
450 
090 
Y=k’ Î S(A)*R(A)*y(A) d A 
450 
090 
Z=k’ J S(A)*R(A)*z(A) d A 
450 
090 
k’=100/5 S ( A )* y ( A ) d A 
450 
f 7 80 
X=k’ j S( A )*R( A )*x( A ) d A 
380 
780 
Y=k’ SS(A)*R(A)*y(A) d A 
3 80 
7 80 
Z=k’ S S ( A )*R( A )*z( A ) d A 
380 
7 80 
.k’=l 00/ S S ( A )* y ( A ) d A 
\ 3 8 0 
X/(X+Y+Z) 
(Hue, Value, Chroma) 
:Y/(X + Y + Z) 
1 
(y, x, ;y) 
A 
1 
1 
! 
a^|A < >Transp. | 
Database-derived 
Empirical model 
4. Verification of the model and summary 
Fig. 8 shows the comparison of the estimated and the observed transparency. Observation 
time lag between the twos is less than 4 hours. The rms error is 67cm and is small enough 
compared with the estimated maximum of 6meters. 
The main points of this study is summarized as follows. 
1) Color of sea database is developed and the model to estimate transparency is derived from the 
database. 
2) Atmospherically corrected Landsat TM data is applied to the model to estimate transparency 
and rms error between the estimated and the observed is 67cm for the maximum estimated 
transparency of 6meters. 
3) It demonstrates the possibility of mapping not only vertical but also horizontal variabilities of 
optical properties of Osaka Bay water based on Landsat TMdata. 
References 
1) Osaka prefectural fisheries experiment station : Annual activity report of Osaka prefectural 
fisheries experiment station, lOOp, 1971.(original in Japanese). 
2) Ota,N.: Colorimetric Engineering. 305pp. Tokyo Denki University, Tokyo, 1993.(original in 
Japanese). 
3) NASA Earth Science and Applications Division : The Highlights of 1989, 78p, 1989. 
4) Chavez,P.S.: An Improved Dark-Object Subtraction Technique for Atmospheric Scattering 
Correction to Multispectral Data, Remote Sensing of Environment, 24, 459-479, 1988.