rary mask
isolate
LICATION
he simple
areas in light
July 1976
ted areas
7 imagery
This proce-
in agricul-
- 174] -
Summary of Application to:
Forest Site Index Mapping in North Central Colorado
Craig Tom Lee D. Miller
Background and Objectives of the Test: Forest site productivity expresses the combined in-
fluence of the biotic, climatic, and edaphic conditions on the timber growing capacity of a
given tract of land. Site productivity is commonly designated by a site index number which
relates tree height to age. That is, site productivity is defined as the average heighth of dom-
inant and codominant crown classes in well-stocked stands at specific ages, usually 50 or 100
years. For example, site index 125 on a 100-year base means that the dominant and codom-
inant trees will average 125 feet or 38 meters in heighth at 100 years of age.
While foresters can manipulate timber density, species composition, quality, and size distribu-
tion through cutting, the potential productivity is determined by the site. However, site
“value” determinations are currently derived through expensive and time-consuming field
measurements of tree height/age relationships. It was hypothesized that the application of
statistical pattern recognition techniques to Landsat multispectral scanner data, combined
with overlays of map-derived ancillary landscape data, would offer particular promise for site
index mapping over large areas.
Study Site: The Eaton Reservoir 1:24,000 scale U.S.G.S. topographic map quadrangle was
designated as the study area for the site productivity mapping study. This test area was im-
aged on the 15 August 1973 Landsat-1 frame designated as path 36, row 32. The site is
situated in the Colorado Front Range approximately 80 kilometers by road northwest of
Fort Collins, Colorado. The site area is rectangular in shape, with dimensions of 10.5 kilom-
eters east-west and 13.8 kilometers north-south, and an area of approximately 145.6 square
kilometers. An abundance of diverse landscapes with a variety of landforms and vegetation
types occur within the small test site. The elevation ranges from 2,341 meters to 2 909
meters above mean sea level. The climate is characteristic of the Colorado Rockies. Abun-
dant sunshine, cool summers with frequent showers, heavy winter snows, low relative
humidity, and wide temperature fluctuations are normal. Annual precipitation ranges from
25 to 38 centimeters, with over half of this precipitation falling during the winter months.
Major tree species include quaking aspen (Populus tremuloides Michx.), Douglas fir (Pseudot-
suga menziesii (mirb.) Franco), limber pine (Pinus flexilis James.), lodgepole pine (Pinus con-
torta), subalpine fir (Abies lasiocarpa (Hook.) Nutt.),and white fir (Abies concolor (Gord. &
Glend.) Lindl.).
Approach and Results: Ground control efforts consisted of initial field inventory and sub-
sequent image registration activities. Field inventory crews of the Colorado State Forest Ser-
vice visited 37 plots within the study area and compiled statistics for timber type, stand size
class, stocking class, site area class, stand area, and site index. The site indeces ranged from
25 through 65 and constitute the training data employed in the subsequent analysis.
The 15 August 1973 Landsat sub-image was geometrically corrected and resampled to yield
137 rows and 102 columns of 1.01 hectare (2.5 acre) square cells for all four MSS bands (Fig.
6.1). A nearest-neighbor algorithm with earth rotation, scan line skew, nonlinear mirror
velocity, frame rotation, and pixel resampling without ground control points was used to
