on and 
1, 
istribution, 
etermi- 
| Wave- 
[o 
U 
0S 0) 
n 
deter- 
The func- 
e scatter- 
iments, 
and « (h) and u(h)are obtained through calculations of à layesed standard 
atmosphere tol using the RADT MO program(?); (h) depends on the 
turbidity of the atmosphere which is estimated from the visibility range. 
Note that «(»o) = 1, pe) zl. 
The equations (2) to (5) are derived in ref. (3). 
Obviously, Tp of eq. (1) is given by eq. (4). The window transmittance, 
Tw of eq. (1), has been determined by Laidet®) for various view angles, 
and has been approximated by: 
C 
“ cos” 
Se N 
  
Note that U is the scan angle with respeci to the normal of the window, 
and is not necessarily equal tc £f. 
Calibration of the OCS data in terms of radiance (mW/sterad cm^nm) 
has posed a me jor problem since calibration values given by NASA. be- 
fore July 1978, were unreliable. Sun glint on a homogeneous water sur- 
face can serve as an interchannel calibration source (Fig.9-12). How- 
ever, that possibility has not been exploited. 
5. hesults 
For flight track 4 (stip track B), the following parameters have been 
correlated with t!: scanner data: sample number (which is highly corre- 
lated with the distance to shore), ralinity, water temperature, chloro- 
phyll A concentration, echo-sounded water depth, turbidity, suspended 
matter concentration detern.ined in laboratory. The correlation coeffi- 
cients were computed both for tide -correctea and not tide-corrected 
data, and for both raw scanner data and for water-leaving radiance, ob- 
tained from the raw data as described before. Since the ship track is 
parallel to the flight line, significant differences in the correlation coef- 
ficients obtained with raw data or with radiances have not been expected 
and indced could not be observed. 
Results are quite different when comparing data corrected for tide and 
the uncorrected data. For both sample number and chlorophyll concen- 
tration, and to a lesser degree for salinity, turbidity, and suspended 
matter concentration, correlation coefficients for channels 2 to 5 are 
closer to 1 (for sample number and salinity -1) for tide-corrected data. 
Note, however, that correlation coefficients between neighbour channels 
(i.e. 1 with 2, 2 with 3, etc.) behave likewise; actually, these coeffi- 
cients are very close to one after tide correction, except for channel