(by remote sensing) and went on to suggest that such an approach may not be 
practical outside coastal waters because it requires a large standing crop of 
phytoplankton. 
If the optical properties of the matter suspended or dissolved in water 
are known then, in principle, one should be able to measure their concentrations 
by remote sensing techniques. The basic ideas are simple. Some of the back- 
scattered underwater light emerges from the water. The spectral characteristics 
of the underwater light should be the same as that of the light which has just 
emerged and which is called the water-leaving light. Therefore, an analysis of 
the water-leaving light and the light inside water bodies should yield the same 
ocean colour. 
The idea of remote measurement of ocean colour in general and chlorophyll 
concentrations in particular did not receive much attention until Clarke et al. 
(3) compared the backscattered light from sea as seen by a low-flying aircraft 
with the simultaneously measured chlorophyll concentrations. Their experimental 
work also concluded that a large fraction of the spectra recorded on an aircraft 
is due to the atmosphere. Gordon (4), Slater (5) and Sturm (6) also showed that 
a large fraction of the remotely sensed signals in the visible spectrum is due 
to the atmospheric effects. Because of the complexity of the atmosphere, the 
difficulty in using the remotely-sensed dataquantitatively lies in the removal 
of the unwanted fraction of the signal.  Reasonably good success along this line, 
due to many investigators (for example, see refs. 4,7-10) led NASA to put the 
Coastal Zone Colour Scanner (CZCS) on board the NIMBUS-7 satellite which was 
launched successfully on 24 October 1978. 
The primary objective of the CZCS was to detect small variations in the 
water-leaving radiance because these variations are caused by the changes in the 
quality of the water. Since the water-leaving radiance is a small fraction of 
the satellite-sensed radiance, the accuracy with which one can extract a useful 
signal from the CZCS data depends on the accuracy of the atmospheric correction 
algorithm. Among many other things, the quality of atmospherically corrected 
data depends on the quality of the inflight calibration constants which will be 
discussed in the next section. 
There are many assumptions involved in the atmospheric correction algor- 
ithm which is being used at present (113. We shall discuss some of these 
assumptions in a separate section and indicate that it is possible to obtain a 
better atmospheric correction algorithm by ignoring some of these assumptions. 
The difference in these two atmospheric correction algorithms will be illustrated 
by presenting a typical example. 
Finally, we shall discuss the uncertainties in the chlorophyll-like pig- 
ment algorithm due to the uncertainty in the solar irradiance in the 443 nm 
channel of the CZCS and some of the oceanic parameters. Using the CZCS data and 
the in-situ measurements, we shall show that the chlorophyll-like pigment algor- 
ithm is too sensitive to even 1% variation in the solar irradiance in the CZCS 
channel 1. 
INFLIGHT CALIBRATION QUALITY 
As remarked earlier, one of the basic requirements for the CZCS detectors 
to be able to detect small changes in the water-leaving signals is that the in- 
flight calibration constants must be so reliable that these small changes in the 
useful signal are not observed. Unfortunately, there have been two types of 
reports with regard to the inflight calibration quality of the CZCS detectors; 
first it has been claimed that the calibration lamp 1 is degenerating (12), and 
secondly it has been claimed that the sensitivity of the CZCS band 1 detector is 
decreasing continuously (12,13). Our interest in the guantitative analysis of the 
CZCS data lead us to investigate along these lines. 
There are two incandescent lamps which are used for the active calibration 
of the first five spectral band detectors of the CZCS. By active calibration we 
mean that during the Earth scan, the detectors view the standard radiance from 
the calibration lamp and the resulting voltage signal is digitized and stored in 
the housekeeping data. Lamp 1 was being used regularly since the successful 
698 
a A 
ee a el