DATA ORGANIZATION 
Before embarking on a project as ambitious as this, it is imperative to have well developed procedures for 
organizing the data. Because the majority of the source maps were in UTM projection, it was decided to use that 
as the base projection for the project. 
The study area, however, spanned two UTM zones, and because the data was far too voluminous to handle as one 
or two very large coverages, it was determined that the preliminary data compilation and digitizing would be done 
on the basis of USGS 1:100,000 quadrangles (1 degree longitude by 30 minute latitude). There are 57 1:100,000 
quads in the study area, and these proved to be a good compromise in terms of overall data flow, coverage size 
and extent, and numbers of coverages to be handled. 
The final coverages are organized into 13 tiles (except for vegetation which is maintained as a single state-wide 
coverage), each approximately corresponding to 4 1:100,000 quads. The actual boundaries of the tiles are the 
public land survey township boundaries closest to the quad boundaries. The township lines provide a more logical 
boundary because of their close correspondence to polygon boundaries in the land ownership data layer. 
One other key element to data organization is the strict adherence to naming conventions. Not only does this make 
it possible track the progress of each portion of the data base, but it enables the use of macros for processing at 
various steps in the data automation process. 
DATA CAPTURE SEQUENCE 
Because of the multitude of data sources and the need to have a data base that would integrate vertically, the 
following sequence was used for data capture: 
1) Import existing digital data 
2) Digitize Public Land Survey data 
3) Digitize Land Ownership from most appropriate source 
4) Fill in and verify Land Ownershsip from alternative sources 
5) Digitize USFS Land Management Plans 
For the data sources that did not contain geographic reference, the public land survey grid served as an intermediate 
reference. The public land survey that had been captured from a controlled source was plotted out at the scale of 
the source maps along with registration tics. The controlled plot was then "best fit" to the source map using the 
public land survey as a reference, and the registration tics were pin pricked onto the uncontrolled source map. The 
source data could then be digitized into its appropriate 1:100,000 quad coverage. 
To facilitate the edge matching process as well as transfer of geographic control to uncontrolled manuscripts, all 
data was captured in real world coordinates, and adjacent features within 1:100,000 quads that had been previously 
digitized were "borrowed" as the digitizing progressed. It was, therefore, essential to keep good records of the 
status of each 1:100,000 quad. The 1:100,000 quad organization also provided an efficient practical size for 
frequent editing sessions. 
The final step in the process was to combine the coverages from each 1:100,000 quad of a given data layer and 
clip them at the tile boundaries. This process was not done until each quad was digitized, quality checked 
internally, sent to the representative land owners for verification, edits made, and a final quality check completed. 
HARDYVARE/SOFTWARE ENVIRONMENT 
Given the time frame (14 months), the size of the project, and the processing and data storage requirements for 
building and maintaining this data base, an NFS network proved to be an ideal computing environment. The NFS 
network included both UNIX workstations (SUN, Data General, and Tektronix) and 386 PC workstations (running 
PC NFS). This environment allowed distribution of the digitizing and preliminary processing on the PCs, and 
consolidation and advanced processing on the UNIX machines.