the surface area of the TU as mapped in the urban
zone to yield the total volume contribution. All the
individual TU contributions were totaled for the
particualar zone. Table 2 shows the results of this
method for the five urban zones, along with the
results from the global method for comparison. This
second method does not have the statistical
Table 2. Estimated fuelwood volumes, total and
average, for the two methods of estimation. Method 1
calculated a global mean volume from the Phase II
field sites placed in the zone, and Method 2
calculates a mean volume for each terrain unit and
then calculates a total for the zone by means of the
surface area of each TU in the zone,
of
METHOD 1 METHOD 2
TOTAL
AVERAGE
TOTAL
AVERAGE
URBAN
VOLUME
VOLUME
VOLUME
VOLUME
ZONE
(steres) 1
(st/km sq)
(steres)
(st/km sq)
Niamey
27,008,082*
860.13
13,472,177(a)
556.97
Dosso
29,539,555*
940.75
17,271,113(b)
712.25
Tahoua
5,810,643*
185.05
7,099,787(c)
674.74
Maradi
7,328,370+
349.97
6,239,875(d)
317.13
Zinder
11,603,242*
369.53
6,047,305(e)
340.90
* for a complete zone: 31,400 sq km
+ for a partial zone: 21,000 sq km
(a) representing a total area of 24,188 sq km
(b) representing a total area of 24,248 sq km
(c) representing a total area of 10,522 sq km
(d) representing a total area of 19,676 sq km
(e) representing a total area of 17,739 sq km
integrity of the global method because the average
figures were calculated using dat from sites not
necessarily associated with a zone in which the
estimates are developed. It is, however, a method
for providing some check on the results of the
global method. It is apparent from the figures
presented in Table 2 that the second method yields
fuelwood volumes lower than those estimated by the
global method. One reason for this is that, in every
case, the surface area for which TU mean volumes are
available are less than the total area of the zone
for example, in the Niamey zone TU average volume
estimates were only available for 24,188 of the
total 31,400 square kilometers. It is obvious that
these lands for which there are no TU volume
estimates do contribute some fuelwood volumes,
therefore, the total volume figures should be
elevated. However, these lands are generally in the
non-forestry TU categories, and they were not
sampled in the procedure because of the bias against
them. This means that, although there is some volume
contribution, it is likely to be fairly small and
the figures are not overly pessimistic.
4.0 DISCUSSION AND CONCLUSIONS
The primary points of discussion include: 1) the
appropriatness of the data sources used, 2) the
methods applied in using them, and 3) the validity
of the results, both as to their statistical
validity, and, within the context of the stated GON
needs.
4.1 Appropriateness of the data sources used
In general, no problems were encountered associated
with the specific data types used as sources of
information. After suitable training the GON
personnel were able to handle and interpret the
various base datas used, and process the resultant
information. In addition, interpretation fo the base
datas did produce the types and amount of
information required for the base mapping effort,
the sample selection and volume estimation.
4.2 Suitability of the specific methods employed
Review of the implementation of the specific methods
employed in the inventory process indicates that
they were completley suitable form the the standpoint
GON information needs, working condition constraints
and capability of GON forestry personnel. The basic
procedures depended primarily on human effort for
their completion. All of the interpretation
procedures were based on visual procedures, and a
majority of the data handling and analysis was done
by hand. The notable exception is the use of
microcomputer processing to develop the multiple
regression prediction equation and it's use in the
calculation of individual site volumes. Finally, due
to the "pioneer" aspect of this work, emphasis was
placed on documenting the implementation procedures
and interim results so that detailed review is
possible.
4.3 Validity and suitability of the estimated
fuelwood volumes
The validity and suitability of the estimated
fuelwood volumes is affected by two principal
aspects: the application of the regression equation
and the mehtod of predicting volume for each urban
zone.
The development of the prediction equation,
itself, followed accepted procedures, using
appropriately collected paired observations and
standard curve fitting procedures. Some question
remains as to the suitability of using squared and
cubed terms of original variables in the expression.
This is being researched at the present time. The
expression did give suitable results in the Niamey
and Dosso zones when compared with data collected
independently concerning fuelwood volume, However,
in the Tahoua, Maradi and Zinder zones problems were
encountered. The equation was created using paired
observations in the Niamey zone and it’s use under
similar conditions was appropriate (in the Niamey
and Dosso zones). However, vegetation species •
composition, cover percentages, and species utilized
for fuelwood were found to be different in the three
eastern zones. This reduces the suitability of the
application of the equation, and the validity of the
results, from these areas. At the present time
development of region-specific equations is under
way.
Analysis of the results for the individual zones
does show a very high variability. This has been
determined to be a combination of a, by nature,
highly variable resource, and the low sampling
density. Of the two different methods, it appears
that the terrain unit model has the greatest
potential for producing the planning level
information, the global method requiring a much
higher sampling density to appropriately represent
the range of conditions in a zone. Even though the
terrain unit model does not have the statistical
integrity of a system employing independently
selected samples, the inherent capability of
focusing sampling effort is of more value in the
context.
