450 
This provides the ability to efficiently sort and filter 
large amounts of lake data for selective classification and 
reporting of the lakes stored on the file. 
The significant differences between this and past studies 
involve two major aspects. The first was to create a 
computer program that could produce a file that uniquely 
identifies lakes by geographic location (latitude and longi 
tude) for permanent storage and retrieval of lake informa 
tion. The second was to design a system that could use off- 
the-shelf Landsat data to compute accurate map projection 
lake center positions from Landsat scene coordinates. Only 
recently has Ground Control Point (GCP) geometric correction 
been available for Landsat scenes delivered by EROS Data 
Center in Sioux Falls, South Dakota. Although GCP system 
capability exists, operational capability and/or corrected 
data may still be lacking. 
METHODS 
A generalized discussion of the methods used to generate the 
Satellite Lake System (SLS) is provided here to trace system 
development. A "Satellite Lake System Manual" exists 
(Mellor 1982) that describes the system and programs with 
enough detailed documentation both for the user to operate 
the system and for the programmer analyst who may wish to 
make future system modifications. The Landsat Data Users' 
Handbook (USGS 1979) would also be useful to the user and/or 
programmer not thoroughly familiar with Landsat data. Users 
of the SLS need little or no previous experience with com 
puters. The SLS is controlled through a series of key word 
instructions that are accessed upon entering the programs. 
The SLS programs have been designed for and stored in the 
University of Alaska Honeywell Information System (HIS) 
66/20 and are written in FORTRAN and COBOL computer lan 
guages . 
The SLS is comprised of 2 parts: the identification system 
and the retrieval system. Flow diagrams and complete 
documentation are provided in a "Satellite Lake System 
Manual" (Mellor 1982). 
The identification system uses Landsat digital data, the 
driver program LAKEID, and subroutine LCALC to generate a 
master lake file. Each lake is identified by latitude and 
longitude of its centroid and is characterized by area, 
perimeter, and crenulation calculations. In addition, the 
identified lakes are plotted on an indexed lake finder 
output that can be used to verify a lake's shape and size in 
relation to the geographic context of other surrounding 
lakes. Another program (UPLAKE) is used to input additional 
data into the master lake file and to generate a listing of 
updated lake information in a printout that verifies the 
master file modification. 
The retrieval system program RETRIEV uses the master lake 
file to generate the following optional listings: 1) a 
catalog of all lakes on file in order of descending latitude 
and longtitude; 2) a catalog of all lakes on file sorted by