152 
Essentially, the cumulative distribution functions described above, cont 
which can be generated for any combination of wildland units, specify imag 
the portion of the south one-half of the Prescott National Forest that part 
supports arbitrarily defined minimum basal area levels. The development redu 
of this statistic satisfies the primary objective of this study. 
RESULTS 
Forest stocking equations were developed from the source data generated syst 
by implementing the methodology outlined above. Since the primary sampling meet 
units were selected by random sampling, it was assumed that the assessment on a 
of the ground estimates of basal area was unbiased. Thus, accepting this stoc 
assumption, the portion of the south one-half of the Prescott National curr 
Forest that supports arbitrarily defined minimum basal area levels can the 
be estimated from the portion of the primary sampling units that supports sequ 
corresponding minimum basal area levels. to t 
To mathematically derive the forest stocking equations, the cumulative 
distribution functions representing ground estimates of basal area that unif 
were obtained, in part, through the interpretation of 1:120000 imagery of n 
with four levels of magnification were subjected to regression analyses. deve 
Only the equation involving the 25X magnification will be presented: of t 
deci 
(7) Y = 1690 - 42.7(X) + 0.272 (X 2 ) the 
f avo 
r = 0.98 fore 
that 
where Y = the portion of the south one-half of the Prescott National fore 
Forest that supports minimum basal area levels, in percent prod 
to m 
X = minimum basal area level, in square feet per acre. Obvi 
maxi 
r = correlation coefficient the 
Solving forest stocking equations for numerous alternative minimum 
basal area levels can be time consuming. Therefore, supplementary computer to a 
programs can be written to solve equations in terms of the portion of a eval 
wildland area that supports any intermediate minimum basal area level. 
Estimates obtained from a graphical presentation of an equation will often fore 
suffice, however. melt 
land 
Various types of errors of known and unknown magnitudes were associated cons 
with the development of the forest stocking equations described herein. mana 
While many of these errors were considered to be generally linked to many impl 
remote sensing exercises, one problem was somewhat unique to this study. what 
a pr 
In the analysis of crown closure on the 1:120000 imagery, large areas "Wha 
of dark shadow within the primary sampling units frequently made it difficult prof 
to discern shadow, ground, and forest crown closure. Possibly, if more