will be discussed in this section. The results presented in this section are 
based on the correlation of in-situ measurements with the CZCS data for the east 
coast of Scotland; a complete report on this work is in preparation. For the 
purpose of present discussion, it should suffice to say that the CZCS data were 
corrected using the OLD atmospheric correction algorithm which utilizes the 
iteration procedure of Smith and Wilson (19). 
Error Due to the Uncertainty in Solar Irradiance 
  
A wide range of variation in the solar irradiance values can be found in 
literature (5,6,11,25). Before the publication of the revised values of the 
solar irradiances (11), it was rather difficult to guess the source of the solar 
irradiances used by many investigators. In fairness, therefore, there should be 
no objection if we say that the solar irradiances used by different investigat- 
ors are likely to be different. Even if this were not true, the revised values 
of the solar irradiances (11) are different, especially for the 443nm channel of 
the CZCS. Since the chlorophylls have a strong absorption band in this spectral 
channel, a slight change in the value of solar irradiance in this spectral 
channel could have an appreciable effect on the chlorophyll-like pigment algor- 
ithm. For this reason we have examined the effects of the variation in the value 
of solar irradiance E(443nm) in the CZCS channel 1 on the parameters a and b in 
the chlorophyll-like pigment (Chl) algorithm 
Chl = a(L,(443nm)/ L, (550m) )^ (5) 
where L,(A) stands for the water-leaving radiance in the spectral channel A. The 
results ane summarized in Table 2 where R is the correlation coefficient. This 
table shows that the parameters a and b in Eqn.(5) depend very strongly on the 
value of E(443nm). It is easy to see from this table that even 1% difference 
in the value of E(443nm) can alter the parameters a and b to a large extent. 
Since the revised values of solar irradiances differ (11) from the values used 
previously, the data in Table 2 suggest that all the previously reported results 
should be revised using the revised values of the solar irradiances. 
  
Table 2 
Chlorophyll-like pigment algorithm as a function of solar 
irradiance E(443nm) on the top of the atmosphere. 
  
  
E(443nm) Correlation - 
mW Coefficient b 
cm —um R (wg/1) 
180 -0.973 4.487 -7.059 
182 -0.939 3.097 -7.198 
184 -0.940 2.070 -7.361 
186 -0.941 1.329 -7.554 
188 -0.942 0.813 -7.784 
190 -0.943 0.468 -8.063 
192 -0.944 0.250 -8.401 
194 -0.945 Q.121 -8.817 
196 -0.944 0.051 -9.335 
198 -0.943 0.018 -9.983 
  
Uncertainty Due to Salinity and Temperature Variation 
The refractive index of sea-water, which appears in the atmospheric 
correction algorithm, depends on the salinity and temperature of the water. 
Since the salinity and temperature of the sea-water vary from place to place and 
from time to time, it is not possible to include the variations in these para- 
meters in the atmospheric correction program. But, one can estimate the 
704 
nm» N WA 4 xx 
ERE oT a rennes 
a nn 
unce 
int 
lies 
chan 
conc 
of S 
Jörg 
intr 
5%. 
nifi 
radi 
wate 
ianc 
pigm 
pher 
ment 
this 
temp 
refr 
that 
abou 
This 
Coun 
11, 
12, 
13, 
<A