832
PROVISIONAL SELECTIVE ELIMINATION KEY
TO FIVE CROP TYPES ON THE NORTHERN PLAINS
1. DENSITY VALUE BELOW .58 ON LATE SPRING* IMAGE;
BELOW .50 ON LATE SUMMER** IMAGE 2
1. DENSITY VALUE RANGE .76 THROUGH .86 ON LATE
SPRING IMAGE; .77 THROUGH .95 ON LATE SUMMER
IMAGE Summer F allow
2. DENSITY VALUE RANGE .59 THROUGH .84 ON
LATE SPRING IMAGE; .36 THROUGH .47 ON
LATE SUMMER IMAGE Sod
2. DENSITY VALUE BELOW .52 ON LATE SPRING
IMAGE; BELOW .36 ON LATE SUMMER IMAGE 3
3. DENSITY VALUE RANGE .34 THROUGH .48 ON LATE
SPRING IMAGE; .12 THROUGH .26 ON LATE SUMMER
IMAGE Spring Wheat
3. DENSITY VALUE RANGE .27 THROUGH .39 ON LATE
SPRING IMAGE; .20 THROUGH .32 ON LATE SUMMER
IMAGE Oats
3. DENSITY VALUE RANGE .44 THROUGH .52 ON LATE
SPRING IMAGE; .24 THROUGH .36 ON LATE SUMMER
IMAGE Barley
*1000 Degree-Days
**2000 Degree-Days
FIGURE 3. The results of using the data shown in Figure
2 to construct a selective elimination key.
season progress. For a detailed discussion of the degree day concept as
used here, see Nuttonson (1957).)
In order to correct this deficiency in the key, a new comparison (p
matrix must be constructed which provides an exclusive tone range for pr
each crop. Figure 1 is inspected in order to select cut-off points for
the tone ranges of each crop type. It is this mandatory procedure which
imposes the most serious transgressions in information transfer between fo
raw descriptive data and the final version of a selective elimination key. da
The revised comparison matrix is shown in Figure 4. A selective elimina- th
tion key can be constructed using this form of data which would be of
exhaustive, i.e., there would be only one possible crop type for each set va
of tone values. In order to construct such a key, however, valuable
diagnostic information would need to be sacrificed. (Those tones which
must be "sliced" from each crop type in order to eliminate overlap between ea
tone ranges can, therefore, no longer contribute to the identification ac
process).
co
th
Fi
th
ap
ca
Thi
a 1
ti'