fuelwood volume categories. The conditions which met 
these criteria were: land form, vegetation type and 
soil type, and the maps produced were called soil/ 
vegetation/land form (S/V/LF) base maps. 
The initital step in the production of the S/V/LF 
base map was to divide each urban zone into sub 
zones based on their LANDSAT image characteristics. 
The LANDSAT image was used at this initial level, 
even in spite of it's low resolution, because the 
base data was generally consistent for the entire 
urban zone, ie., a characteristic reflectance type 
was likely to represent the same ground conditions 
anywhere in the zone. These sub-zones, called 
"cartographic units" (CU's), were, generally, 
greater than 25 square kilometers in area and were 
characterized by several different LANDSAT 
reflectance types. The next step is to define the 
relationships between LANDSAT reflectance types and 
ground conditions, and to evaluate the correlation 
between these identifiable ground conditions and 
fuelwood volume categories. 
Field sites were located in each of the different 
LANDSAT reflectance types in order to determine 
their corresponding ground conditions. Two to three 
sites were placed in each major reflectance type in 
order to determine the variability of ground 
conditions represented. Evaluation of this 
variability showed that for several major 
reflectance types the variability in predicting 
ground conditions was great enough to reconsider 
using the LANDSAT imagery alone as the base data for 
extrapolating the site-specific data to the entire 
zone. It was decided, therefore, to interpret the 
IGN aerial photography to determine the ground 
condition present in each of the CU's of a zone. 
Representative portions of each CU were identified 
and their aerial photo coverage assembled. The 
photos were interpreted in stereo format using a 
mirror stereoscope. The recognition of specific 
ground conditions, correlated to fuelwood volume 
categories, has been standardized through the 
application of a photo interpretation key. This key 
was developed in the process of defining a series of 
ground condition categories, called "terrain units" 
(TU's). These units, which are distinctly different 
and are characterized by having homogeneous 
conditions of landform, soils and vegetation species 
composition, have been established by analysis of 
data from more than 1,200 sites placed in the field 
in Niger. Each of the individual TU's is represented 
by: 1) verbal descriptions of their characteristic 
landform and terrain conditions, soils 
characteristics, and vegetation species 
compositions, and 2) a series of stereograms made 
from the IGN aerial photography showing the 
appearances of the TU. The result of the 
interpretation process is a map of the selected area 
in a particular CU showing the distribution and 
extent of each TU observed. The surface area of each 
TU in the interpreted zone is determined using a 
polar planimeter and it's relative area calculated. 
The results from this interpreted zone are 
considered representative of the entire CU and a 
legend marked within the CU boundary, as shown in 
Figure 2. This figure shows the identifying number 
of each CU, it's surface area, the codes 
representing the TU's interpreted, and their 
percentage surface area (the number in parentheses) 
in the CU. Each CU in the urban zone is treated in 
this manner until the S/V/LF base map is completed. 
5.2 Phase II: Sampling procedure and estimation of 
ruelwood volumes for the five urban zones 
J.2.1 Selection of field sites for the collection 
of ground data 
With no data available concerning actual volume 
levels or their variability, assumptions had to be 
made regarding appropriate sample size, method of 
allocation and data analysis to produce the volume 
estimates. The number, and method of selection, of 
the field sites was determined based on the eventual 
use of the information, the constraints, and the 
data bases from which information could be 
extracted. The eventual use of the information, in a 
national level planning context, indicated an 
appropriate level of effort to be expended in the 
collection of field data. With no comparative 
information available, and because the needs were 
only for very basic and "first look" types, neither 
accuracy, nor precision, of the quantitative results 
were specified by the GON. The number of field 
data collection sites was defined by the time 
available for field work, and the capacity of the 
RIM section for collecting site-specific ground 
data. Finally, the specific method of selecting 
ground site positions was determined by the 
availability of: the LANDSAT MSS images, the S/V/LF 
base maps, the IGN topographic maps, and the IGN 
aerial photography. 
For a complete urban zone, with a surface area of 
31,400 square kilometers, a total of 75 field sites 
would be selected. The RIM section field crews 
complete an average of three sites per day when 
collecting quantitative data. The average working 
month is 20-25 days, which defines the total of a 
maximum of 75 sites. It must be noted that, the 
Maradi zone is only 21,000 square kilometers and 
only 50 sites were selected. The total number of 
sites completed in the field, at which quantitative 
data was collected, was 349. 
The method used to select the individual sites 
within each zone is one which is classically applied 
where several levels of base data are available 
(especially remotely sensed data). This is a 
multi-stage, or multi-level approach where the full 
zone is divided into sub-zones of equal area by 
interpreting the most general base data, and a small 
portion of these selected for more detailed analysis 
using the next most detailed base data. This 
effectively eliminates a large portion of the zone 
for the eventual placement of field sites. These 
sub-zones are again divided into equal parts and 
again a small proportion selected for more detailed 
examination. For each zone the LANDSAT was 
Figure 2. A portion of the soil/vegetation/land form 
(S/V/LF) base map showing the lines defining the 
cartographic units (CU's), and the legend showing 
the CU number, it's surface area, the terrain unit 
codes, and their respective surface areas in 
die CU.