in which M total number of PU's in population 
N, = total number of SU's in pu; 
  
  
  
  
J 
2 2 k 3 : : 
sy = s within = variance within PU's 
2 
m n, 
mn 5 J 
# 2 2.12 yi 
S yx. à 
ij n. 
2 J 
s within = 
m (n, - 1) 
J 
2 2 
2 S between - s within 
53 
n, 
J 
m n 
2 m n 
=, | - =, 
- ij y S. ij 
2 p t = n 
S etween n - 1) 
2 : 
(s between = variance between PU's). 
Unequal probability sampling, such as sampling with probability proportional 
to size or to estimated size (PPS) is most effectively used for PU's of 
unequal size. Then different formulae should be applied and the mean volume 
is then written as: 
m 
Pd. 
m x 
J 
» 
and the variance: 
m y 2 
v la. 
- 1 my Y, ]2 47x. 
v = 4 :[-] 215,1] 
(m-1) | — x, m 
In these last two formulae x, = area of forest per PU, 
= volume of forest per PU 
ys 
  
The probability of selecting PU, , applying PPS is p, = — 
J 
so that for the total volume: 
  
     
Vy 
When 
quot: 
For 1 
and 
(Loe: 
Also 
The : 
appl: 
whic] 
with 
1:10 
scal 
each 
fore 
of tj 
and |