June 1976 Landsat Image Depicting Three APU's Located 1n
the Texas and Oklahoma Panhandles
Figure 3.
throughout the Great Plains. Crop Reporting Districts (CRD's)
were used as the sampling base. These districts are political
subdivisions and do not reflect true differences in production
of agricultural commodities d::» to natural phenomenon. Each
segment was analyzed during the crop year to enable estima-
tion of Great Plains wheat production. The 601 sample segments
were an oversample which necessitated the use of additional
time and resources for data analysis.
In 1976, the Great Plains area was resampled using APU's as
the sampling base. The number of sample segments decreased
to 487, which reduced the data load by 23 percent.
b. Improved Yield Homogeneity. During 1975 and 1976, yield
computations were based on CRD's. This inherently resulted
in large variances for yield estimation. A study was con-
ducted in Kansas to determine if APU's increased yield
homogeneity. County yields for each partition were obtained
over a 19-year period, starting with 1956. The variability of
county yields without APU's was 142 percent higher than that
within APU's, indicating that APU's do divide the state into
regions where county yields are homogeneous. An analysis
was done to compare the gain in efficiency between APU's and
CRD's. APU's had a relative gain of 41 percent, which shows
that APU's are more homogeneous than CRD's.
c. Mheat/Small Grain Confusion. A major problem in the
LACIE is separating wheat from other small grains. When
analyzing a sample segment, it is extremely valuable to have
cognizance of which small grains are grown in the area and
whether or not wheat is the major crop.
Each APU was categorized according to the percent wheat,
type of wheat, and the other small grains. Oklahoma is
selected to illustrate the results. Table ] lists the
information applicable to the APU's in Oklahoma.
Figure 4 is a map of Oklahoma which contains country, CRD,
and the APU 7 boundary lines. The percent wheat is plotted
for several counties. Note the drastic differences in yields
for those counties which are separated by the APU boundary
lines. These lines were develcped along natural boundaries
as discussed in the development of APU's. From the LACIE
Stnadpoint, the area within APU 7 has similar growing con-
ditions, thus making it easier to analyze any sample segment
falling within this APU. This would not be the case if this
area was subdivided along political subdivisions.