Page 13 - Interp Book

in the classification (step 3). Stream buffers and ecoregions were also
considered abiotic variables for modeling (see step #7).
5. Development of image objects (segmentation) from air photos at 10 m resolution.
Input data for object creation was from a single variable generated from 3-band
air photos (see methods, below).
6. Attribution of image objects with land use/landcover (from step #3) and abiotic
variables (from step #4). The attribution of small image objects with information
from several separate data sources, each with their own spatial resolution and
source, was an involved process that required tiling and systematic application of
rules for dealing with issues.
7. Development and application of an Ecological Systems and Subsystems (current
vegetation) classification and mapping models based on the attributes assigned
to image objects in step #6. This was an iterative process that required on-
screen viewing and revision of results. The end result of this step was a draft
ecological systems map.
8. Rapid field survey and collection of strategic, georectified data points related to
issues discovered in the draft map, and revision of the map.
9. Development of final map and database. This step involved use of ancillary data,
such as roads, to improve the look of the map.
10. Development of interpretive materials and final geodatabase. This involved
modification and additions to technical descriptions of ecological systems and
mapping subsystems that corresponded with mapped types, plus development of
new, short descriptions of mapped types. This step also resulted in summaries
and caveats for users.

5