Interdependencies of radiation quantities as functions of time of 
day are presented in the next two graphs. Figure 11 presents the ratio 
of the total downward irradiance at an altitude of 5 km to the total 
irradiance at the Earth's surface. We see that the irradiance at the 
aircraft is slightly greater than on the ground, and becomes more so 
with an increase in haze below the aircraft. There is relatively little 
change in the ratio for two hours either side of solar noon when 
conditions are clear (V-23 km), but the time dependence for that interval 
increases for shorter visual ranges. In all cases, there is strong time 
dependence during early morning and later afternoon hours. This type of 
information can be helpful in using sky sensor signals for signature 
extension. 
The second interdependence, presented in Figure 12, is the ratio 
of path radiance to total radiance. These curves clearly show that path 
radiance is a large fraction of the total radiance, especially for very 
hazy conditions. The important point of Figure 12 is that, for reasonably 
clear days, the ratio of path radiance to total radiance is essentially 
constant for several hours at mid-day for a given atmospheric condition 
and a fixed surface albedo. This fact should be of value in the develop- 
ment of techniques to remove atmospheric effects from data. It is worth 
noting that the difference between atmospheric effects at an altitude 
of 5 km and at low altitude is greater than between 5 km and space 
altitudes. 
Additional models for thermal diurnal temperature predictions for 
various kinds of objects have been developed [23] and applied to the 
detection of Florida sinkholes [24]. Oil slick detection models [25] 
and water depth measurement models [26] have also been developed and 
used successfully to devise automatic processing techniques. 
3. Automatic Techniques and Methods 
In this section I will discuss automatic processing functions, the 
techniques and methods used to achieve these functions, and processing 
equipment. 
A particular point of view must be preserved: that of the informa- 
tion needs of the problem-oriented user. The fact that the; user of 
such data is not directly concerned with the techniques or physical 
parameters of sensing and processing but is, rather, interested in mapping, 
identifying and studying specific objects on the surface of the earth 
and the interrelationships of these objects, must be foremost in the 
conception and implementation of multispectral resource information 
systems. The user would like to know, for example, how many bushels of 
wheat and corn will be produced in a county of so many acres, how many 
ducks will be successfully raised in the prairies, how much water is 
needed to preserve the Everglades and where are and what are the various 
sources of pollution for the rivers, lakes, and oceans.