The
1977).
ween
me is
'thods
. not yet
le two-
lated
| for
as much
are more
day units
- 1827 -
The figures of Table 5 indicate that the difference in office man-day
units between the unrestricted and the two-strata sampling method is
very small.
The gain in the number of field days in the stratified sampling method is
outweighed by the extra office man-days. Of course the difference would
be greater for a higher number of plots, which would occur for a standard
error of the mean smaller than 5%.
Two-phase sampling seems to be most eificient. A drawback of this method,
however, is the correct location in the field of the randomly pricked
photo plots, which requires a photographic scale between 1:5,000 and
1:10,000 for stands of > 41 years.
For two-stage sampling a scale smaller than 1:10,000 can be used. When,
in addition, selection of secondary plots is done systematically, this
method, in spite of the higher office man-day units, may become more
interesting than the two-phase sampling.
In the example, extra photographs at a scale of 1:10,000 (the smallest
allowable scale) are already used for the two-phase sampling in addition
to the 1:30,000 scale photographs. Flying 1:30,000 scale photographs and
constructing 1:10,000 scale orthophoto mosaics is cheaper than the
construction of 1:10,000 scale orthophoto mosaics from 1:10,000 scale
photographs. The difference in cost for the described area including
expenses for flying, photo material, aerial triangulation, ground control
and detail transfer, amounts to approximately 12,200 US$.
Therefore, considering only the additional flying plus the photographic
material costs for 1:10,000 scale photography, the extra expenditure
approximates $1476.- which is equivalent to about 18.5 man-days (at
$10/man hour).
For the two-phase sampling, instead of 16.8 (Table 5), now 35.3 office
man-day units are needed, which is higher than the 25.5 (Table 5) as
calculated for the two-stage PPS sampling.
The advantage, expressed in man-days, of two-stage PPS sampling over
unrestricted sampling equals 31 per cent. The advantage Of two-phase
sampling over unrestricted sampling reduced from 37 to 5.5 per cent after
considering the additional expenses for the 1:10,000 scale photographs.
When, in addition, the area is als0 larger in size than 10,000 ha, it will
be more unlikely that the 1:30,000 and the 1:10,000 scales are flown in
one day. In this case, costs must be added for addittonal flying from the
airbase to the target area, double trip, at a rate of $400/hr.
The resulting additional cost or number of man-days then depends on the
distance from airbase to target area.
References
D.A. Stellingwerf, Results of two-stage unequal probability (PPS)
1978 sampling for timber volume using an orthophoto
mosaic. (Comm.VII-ISP-Freiburg 1978; IUFRO, SG.6.05
Freiburg 1978)
F. Loetscn, Z. Zührer, Forest inventory, vol. II, 469 pp.
K.E. Haller,
1973
Trân-Châu-Lam and Efficiency of several sampling methods
A.A. Khan, (M.Sc. Studies ITC)
1977
M.H. Hansen, W.N. Hurwitz, Sampling; survey methods and theory, vol. 1.
W.G. Madow, Methods and applications, 639 p.
1953