149 
iphic, 
• lish a 
' lead 
latural 
to ’'match" 
(Otentially 
;nt 
>n and 
mits 
iriority 
irizona, 
specified 
;nt 
and, 
in be 
mal 
•pulations, 
:h may 
mtified 
ystem. 
:d, do 
“istics, 
;. In- 
t that 
! associated 
to: 
;uit- 
sh 
statistic 
i may be 
minimum 
ey, 
!St 
!, etc. 
l obtained 
(Ffolliott 
•est 
ie forest 
types of Arizona that support arbitrarily defined minimum forest density 
levels by applications of remote sensing techniques and utilizing high 
altitude photography is outlined. Coupled with satisfying the stated 
general objective, forest stocking equations may also allow for assess 
ments of current and future management status by identifying existing wild 
land units available for possible implementation of given management 
systems. 
In addition to providing input potentially necessary for rational 
decision-making in Arizona, the study yields general methodologies suit 
able for the development of comparable source data elsewhere. 
METHODS 
The synthesis of ground data for the development of forest stocking 
equations involves: (1) the assessment of forest density conditions on 
primary sampling units from high altitude 1:120000 imagery taken by U-2 
aircraft, and (2) the translation of these assessments to ground estimates 
of forest density conditions using standard 1:15840 imagery as an inter 
mediate adjustment basis. 
Basal area, defined as the total cross-sectional area of all tree 
boles on a per unit-area basis, was used as the expression of forest density. 
Basal area was chosen because it has been found to be consistent, easily 
determined, and widely used. 
To illustrate methodologies, stocking equations were synthesized from 
imagery of the forested portion of the south one-half of the Prescott 
National Forest in central Arizona. This study area is representative of 
the cutover ponderosa pine forests that are common to the southwestern 
United States. 
Analysis of 1:15840 Imagery 
The existing US Public Land Survey network was employed to delineate 
primary sampling units of the 1:15840 imagery. A primary sampling unit, 
by definition in this study, was one-square-mile (approximately 259 
hectares) of land which was forested over one-half of its areal extent. 
There were 141 primary sampling units (N) on the study area. To 
select a sample for investigation, an identification number was assigned 
to each unit, and 50 primary sampling units (n) were randomly drawn. The 
selected sample was then delineated on the 1:15840 imagery, using flight 
line maps obtained from the USDA Forest Service to identify locations. 
Two variables, average crown closure of the dominant forest stand 
and average height of trees in the dominant stand, were measured to 
determine forest density conditions. Both of these variables are fre 
quently used in aerial forest density tables (Moessner, 1964).