whi 
  
It must be emphasized that evaluation of an APU approach is 
not a one-time event, but a year-to-year continuing effort. pon 
This aspect of the approach is, in part, what has led the des 
authors to suggest maintenance of APU's via an automated ce 
technique. The update mechanism must depend upon a «n 
relatively unbiased data source and an efficient technique in 
to apply that source. This latter criteria highlights the is 
value of satellite/aircraft color infrared imagery such as eve 
Landsat. The remainder of the paper will look at APU's as 
input to an automated data base and their applicability via Wit 
the data base to remote sensing applications. sup 
| 
III. APU's as a Basis for Automated Data Base Design bed 
dev 
A. General app 
The exploitation and. interpretation of remotely sensed, multispectral, B 
and meteorological data rely heavily upon the utilization of automated 
techniques to achieve cost efficiencies. A potential user of these 
data sources has been, is, and will be faced with a classic system 
design problem of attempting to achieve responsiveness to user needs, 
flexibility in design, and timely throughput at the least possible 
cost. In recent years, designers of information handling systems 
have realized that a major contributor to design, maintenance, and 
conversion costs was the application support data base(s).5 The 
data base and its design is also a major contributor to the degree 
a System is responsive to user needs and to the extent of flexibility 
within a total system.9 The data base viewed in this context becomes 
a critical design consideration when manipulating massive volumes of 
multispectral, meteorological, and agricultural related data necessary 
to support agricultural assessment applications which utilize remote 
sensing technology. The massive data volumes coupled with changing 
technology, changing user demands, and evolving technology create 
an environment which can consume an inordinate amount of resources. 
The USDA Remote Sensing User Requirements Task Force? and the USDA 
Application Test System Design? attempted to address these problems. 
Our approach to a data base concept evolved as a result of the func- 
tional research and development requirements stated by the LACIE 
and those of USDA management. In summary, these requirements are: 
1) a common data base to support statistical sample allocation, agri- 
culture/non-agriculture delineations, yield, area estimation, and 
production aggregation; 2) a data base responsive to a yearly change 
by management in geographic area and crops to be monitored; and 3) a 
common data base to support analysis of spectral,meteorological, and 
other data for varied users. 
Our previous analysis and development work in defining and quantify- 
ing an APU provided guidance and insight into a data base concept