3.5.5 Extrapolation of the site-specific ground 
data to produce the S/V/LF base maps 
The next step in the mapping procedure is to 
determine the distribution of the ground conditions, 
represented by the TU groupings, in the study area. 
One method of accomplishing this task, especially 
where remotely sensed data is available, is to 
extrapolate from known conditions to those unknown. 
The general procedure is to directly observe 
conditions in a restricted portion of a zone or 
region, and using some form of extrapolation media, 
or implement, deduce conditions in those portions 
not directly observed. In this mapping procedure the 
known conditions are those ground conditions 
observed at each field site, the problem is to 
deduce the distribution of these conditions 
throughout the study area, and there are two 
possible extrapolation implements: the LANDSAT MSS 
data and the IGN medium scale aerial photography. 
When using remotely sensed data as an 
extrapolation implement the process depends on being 
able to find characteistics of the image which 
consistently identify the presence, on the ground 
of a certain set of condition, in this case the 
TU's. In attempting to identify the representative 
image characteristics of ground conditions important 
in terms of forestry, the decision was made that the 
LANDSAT imagery available at the time did not have 
enough consistency in prediction of actual ground 
conditions to be used as the implement for 
extrapolation (however, it was decided to retain the 
initial statification, done at the level of the 
LANDSAT imagery, as the reporting base). This 
decision left the IGN aerial photography as the 
extrapolation implement. Given the decision to use 
the IGN aerial photography, it was necessary to 
develop a very basic interpretation aid to 
standardize the process. The type of aid which was 
selected was a photo interpretation key. 
3.5.5.1 The photo interpretation key 
The photo interpretation key was developed 
primarily in response to the need to interpret the 
IGN aerial photography as the extrapolation 
implement. However, other aspects did enter into the 
consideration which include: the need to 
systematize the extrapolation process, the desire to 
emphasize human interpretation or analysis of data, 
the level of training and experience of the Nigerien 
forestry personnel, and the desire to initiate a 
system which can be flexible and applied in other 
areas. 
The key was constructed in a standard manner using 
examples from the IGN aerial photography to be 
interpreted. The photo appearances of each of the 
field sites used in the development of the TU's were 
again examined using a mirror stereoscope in order 
to ascertain what constituted "typical" photo 
appearances for each of the TU series. Once this 
"typical" signature was established it was also 
possible to identify those photo appearances which 
were different from the "typical" signature, but 
still indicated the same TU. This process defined 
the acceptable range of photo appearances for each 
TU. 
A series of stereogram examples, illustrating 
"typical", and acceptable variant, signatures were 
constructed in a format suitable for viewing with a 
pocket stereoscope. Each stereogram was mounted on 
card stock paper, a clear sheet of acetate was 
hinged to cover the photo pair and the subject 
terrain unit indicated in fine ink on the overlay. 
This permitted the interpreter to use the overlay to 
identify the subject TU, or remove it for 
unrestricted viewing of the stereogram. Another 
sheet of clear acetate is placed over the pair and 
an interpretation completed which shows the subject 
TU in relation to it's normally associated 
series. These interpretation results were then 
transferred to the card stock in a position which 
allows rapid comparison of the photo appearances and 
mapped results. An example of a "typical" signature 
for the TU series 841 is presented in Figure 3. 
3.5.5.2 Use of the photo interpretation key in the 
extrapolation process: application in the test area 
A complete set of stereograms has been produced 
for the 52 terrain unit series. There are in excess 
of 200 individual stereograms and associated 
interpretation results. Interpreters were given the 
full set of stereograms and the accompanying TU 
descriptions (available in both english and french) 
to study prior to starting the interpretation. The 
interpretation proceeded by first identifying where 
the photo interpretation was to be done. The two 
arrondissements, Mirria and Matameye, were divided 
into cartographic units, and zones within each 
appropriate CU defined for interpretation, using the 
aerial photography. The photos providing stereo 
coverage of these zones were identified and 
assembled. A plastic overlay was developed with a 
permanent square etched on it covering a 10km x 10km 
block at a scale appropriate to the photography. 
This was overlaid on the middle photo of the stereo 
triplet and the position of the square marked on the 
film containing the CU stratification. This zone was 
numbered and the aerial photos used, noted. 
The aerial photos were then interpreted using a 
mirror stereoscope. With the aid of the photo 
interpretation key the zone was delineated and each 
TU series occurance noted on the overlay. An example 
of such an interpreted zone is shown in Figure 4. In 
the figure it can be seen that there is a heavier 
line which almost cuts the zone in half. This line 
represents the boundary between two CU's, and the 
interpretation results were analysed separately. 
Then for the zone, or for each CU if there were two 
or more, the relative surface areas of each 
identified TU was determined using a compensating 
polar planimeter. These percentages were then 
considered to be representative for the entire CU 
and notation was made on the initial stratification. 
In some instances, especially where there were very 
large CU's, two or more photo interpretation results 
were available. When this occured the results were 
pooled and the CU map code represented an average of 
the observations. 
This process was continued until all CU's 
considered appropriate were evaluated and TU 
percentage codes were generated. In some instances 
CU's were not subjected to photo interpretation 
because their LANDSAT reflectance so closely 
resembled that of another, which had been so 
evaluated, that the results were extrapolated. In 
this manner the initial breakdown was completed 
using a regional type data base, yet the ground 
condition information presented on the map contained 
a level of detail and precision originating from the 
larger scale aerial photography. 
3.5.5.3 Production of the final soil/vegetation 
land form base map 
The final step is the productions of the S/V/LF 
base map for distribution to the users. Using the 
results of the initial stratification the CU's were 
numbered consecutively and each planimetered to 
determine it's surface area. These were both noted 
on the final map sheet. The TU codes, and their 
respective relative percentages, were then added and 
the map drafted in an appropriate format. The most 
efficient format proved to be the base map with the 
addition of major and secondary roads and principal