Previous Page Table of Contents Next Page

Test for the study of the evolution of vegetation in the Sahelian zone

Pierre LEROUX (*)

(*) P. Leroux: Responsable de la Division de Télédétection, B.D.P.A., 202, rue de la Croix-Nivert, 75738 Paris Cedex, France.



This test was carried out in 1971 as a part of a more general study of the whole of the inland delta of the Niger River in Mali. The study was funded by the FAO on behalf of the International Desert Locust Control Organisation of Bamako.

It was attempted in this test to extend to the fullest the utilisation of aerial data on an experimental site. From experience, we still have the feeling today that a kind of reticence exists on the part of many concerning the systematic use of new special aerial photographic shots for any number of types of study of the natural environment. Basically this reticence stems not from financial considerations but more from a lack of awareness of these techniques. In addition, the photo library of the National Geographic Institute of Paris possesses millions of aerial photographs which are, as such, "preserved landscapes ", available to everyone; however, this abundant documentation is almost never consulted.

In the first instance, the object of this test is to review the traditional techniques for the treatment of this kind of information. Great progress has been made since 1971 in the gathering and treatment of aero-spatial information. For example this study would not be carried out today with the same emulsions.

In our opinion, it is necessary for the specialists of the various disciplines involved with the natural environment to keep up-to-date with regard to these techniques in order to extend their use.

With regard to inventions and cartography, these are absolutely indispensable.

The outline of the study

The Sagoumane site is located in Mali, 30 km south of Dioura on the western periphery of the inland delta of the Niger.

Several technical and practical considerations governed the choice of this test zone. From the logistic point of view, the Dioura air strip repaired for our use, permits good liaison with the base at Kara, 30 minutes' flying time away. Housing is facilitated by the presence of an IDLCO settlement. and several tracks make field work relatively easy.

On the technical side, this site was chosen after several low altitude flights over it. Independent determining factors were: the considerable evolution of the western area; the paucity of vegetation data on the standard map at 1/200,000 of the National Geographical Institute; and as the final important factor, the site was photographed in 1952 at 1/50,000 with panchromatic emulsion and in 1971, for the UNDP project, at the same scale but with double panchromatic emulsion.

These combined factors made the Sagoumane site a fairly representative study framework in spite of its relatively limited size.

Field work

Field work consisted of as complete a reconnaissance as possible for the preparation of the photo-interpretation work, which was to be carried out in Europe on the return of the mission. The documents presented hereafter were prepared under these conditions, with the normal process of exploitation.

In respect of the vegetation, reconnaissance on the ground generally confirmed the impressions gathered in the flights.

All the known plant formations of the region are represented on the Sagoumane site: shrubby, dense, and open savannahs dominate and are present in every stage of deterioration.

Difficulties occur in field work as soon as one attempts to distinguish between the cultivation of a year, the fallow of a year, and the fallow of a true grassland savannah. Such distinctions are even more delicate on aerial photographs. The answer rests with the plant ecologist rather than with specialists in our field. Nevertheless, on the whole, the impression gained from the field work and from the close study of the photos allows us to consider the dense shrubby savannah as the original plant formation of the greater part of the study area, the other formations presenting an appearance either of deterioration or of slow recovery.

The examination of photographs was to later provide answers concerning both the recovery and the apparently irreversible destruction of this basic vegetation. At the time of the field work, these answers were not known, and the impressions gained pertaining to the recovery of the basic vegetation were very pessimistic.

On this site, the comparison of the panchromatic and infra-red photos provided nothing of interest. In the office, it was not possible to establish the advantage of one emulsion over another, even from a detailed study of the photographs. Practically no marked difference was found on the same negative between one emulsion and another, apart from a slight higher contrast of grey values in the infrared. The panchromatic was a little duller. This quite unusual observation had, in fact, been made in 1970 on former photographic missions on the internal delta.

From this "in situ" reconnaissance, we also brought back proof of a development of cultivation practices over 20 years. Although it was carried on outside the growing season, it was to confirm the gradual disappearance of the "daba" in favour of "ox-drawn cultivation" permitting the clearing and sowing of larger areas.

Examination of the negatives

The interpretation of the photographs was carried out in Paris on return from the mission, i.e. in perfectly normal exploitation conditions. The data were transferred onto a 1:50,000 photographic enlargement of the 1:200,000 map of the Sagoumane site (Map no. 1).

This map was prepared on the basis of the photographic coverage of 1952. Its purposes were essentially topographical. The plant coverage was therefore not the subject of special study, and the map is very poor in this field. Nevertheless, by taking the aerial photographic coverage of that period, fairly detailed documents in terms of vegetation can be obtained.

This last is very important, as it is now possible to do similar work over almost all of French-speaking Africa, which was covered photographically around the 1950s with a view to the preparation of 1:200,000 maps. In general, most of these old negatives may still be obtained from IGN (1).

(1) I.G.N. Photothèque Nationale, 2, avenue Pasteur, 94160 Saint-Mandé.

Exploitation was based on the following technique:

- Preparation, on the basis of a standard legend, of a 1:50,000 document giving topographical information on the area and soil use in 1952 (Map. no. 2).

- Preparation of a second document on the same scale and on the basis of the same standard legend of the 1971 situation (Map no. 3).

- Quantification of the results by means of the plotting of the various classes of soil use on each document.

- By means of comparison, identification of trends and, if possible, drawing of conclusions.

The situation in 1952

The situation is presented in Map no. 2. The document shows clearly two fundamental factors:

- A few hectares cultivated around the two villages of the area, Sagoumane and Tounde;
- The greater part of the area studied is occupied by dense shrubby savannah.

Thorough plotting (2) makes it possible to define areas, expressed in hectares, occupied by each class and from this to establish percentage densities.

(2) In spite of the care used in this operation, the figures which emerged cannot, at this scale, be considered with a coefficient of error of less than ± 10 %.


Areas (in ha)


Dense shrubby savannah



Light shrubby savannah



Grassy savannah



Dense mixed thicket



Light mixed thicket






Flood area*






* Or humid zone.

Dense shrubby savannah

In order to simplify matters, but perhaps wrongly, dense shrubby savannah has been regarded as the original plant facies proper to the site.

This class, which is easily identifiable on aerial negatives, constitutes the largest part of the perimeter under study.

At the end of the work, and with hindsight, it can be seen that the two classes of thicket which it had been thought should be isolated are hardly representative. It is more likely that they correspond to the dense shrubby savannah class. The thickets are small and represent an occupation of only 2.47 percent.

Accordingly, incorporating them into the dense savannah does not alter the conclusions.

Light shabby savannah

Starting with this class, and within the system explained above, we begin the study of facies of deterioration. A part of this plant formation may be regarded as the result of edaphic equilibrium.

In 1952, the light shrubby savannah, with 12 percent of soil occupation, is the second largest class. It is hardly possible to speak of privileged geographical localization, but the block to the east of the village of Tounde is worthy of attention. This part of the sector is the most heavily populated. Indeed, the occupation continues outside the area of study to the north-east. A few crops occur in patches on this block of shrubby savannah. This may be an indication that the light shrubby savannah corresponds to former fallow land in the process of evolving towards the ecotype represented by the dense shrubby savannah.

Grassy savannah

Third in importance with 7 percent occupation, the grassy savannah class is a plant formation which is fairly easily discernible on photographs.

However, it is difficult to place from the evolutionary point of view within the plant ecology.

Part probably comes from edaphic formations which are pure and perfectly stabilized but impossible to isolate in current study conditions. The other part indisputably derives from fairly recent fallow land in the process of evolving towards light, and later dense, shrubby savannah. But of what age are these fallow areas, and in what proportions? It is difficult to answer these questions.

On the site, the geographical localization of a single point attracts attention: around the Niondji pool. There, the grassy savannah constitutes a large block in the form of an inverted Y. Subjectively, the Niondji pool is reminiscent of a village site (1). On the basis of this hypothesis, a former situation may be imagined which explains very schematically the present situation: a village surrounded by crops, exhaustion of the soil to the point that the reconstitution of the ecotype is prevented, abandonment of the village, and disappearance.

(1) Nothing was done during field work but the discovery of broken pottery by turning the soil a little would make it possible to confirm or disprove this hypothesis.

The other grassy savannah areas more or less bring to mind fallow land.


In 1952, crops took up 4 percent of the area. This figure is probably inaccurate, and must be higher than it should be. Experience with aerial photographs and in dry tropical Africa makes it possible to assert that it is very difficult in the dry season not to confuse the crops of the year with the one-year-old fallow areas and sometimes with the two-year-old ones, depending on the ecological conditions in each place.

1971 situation (Map no. 3)

The map showing the situation in February 1971 was prepared using techniques which were basically the same as those for the 1952 situation. The documents are therefore fully comparable.

Nevertheless, the operation of the 1971 photographic mission drew benefit from progress over 20 years due to improvements in lenses, aerial cameras and emulsions used. This is especially notable in the case of detection in more humid areas. The images are clearer, more accurate; there is more contrast; in short, they are of much better quality than in 1952.

The following observations are made without taking account of the 1952 situation

In 1971, the perimeter may be divided into two almost equal parts:

- The western part seems free of any human interference, apart from the tracks.
- The eastern part is cultivated, and is a mosaic of various facies of soil occupation.

Plotting shows the following distribution as between the various classes:


Areas (in ha)


Dense shrubby savannah



Light shrubby savannah



Grassy savannah



Dense mixed thicket



Light mixed thicket






Flood area






In topographical terms, the network of paths is dense in the eastern part.

The presence of several livestock pens and some cultivated plots surrounded by enclosures made of dead thorn bush attest to the existence of domestic animals in the study area.

Dense shrubby savannah

This plant formation, with 58 percent occupation, is the largest. It is found mainly on the western half of the study area.

The thicket formation, which is relatively poor and unrepresentative of the sector, may be classed with it.

On the eastern half, the dense shrubby savannah, in mosaic form, is clearly in the process of disappearing, as it is very heavily penetrated by crops.

Light shrubby savannah

This class takes second position, with 15 percent occupation, spread throughout the site, without a privileged geographical location except towards the south-east of the perimeter under study. A relatively large block partially encircles the village of Sagoumane to the north and west.

Grassy savannah

The grassy stratum occupies 10 percent of the perimeter, and stands in fourth position. The large blocks are found around the Niondji pool and to the west of Sagoumane. The remainder is spread in small areas throughout the site.

Mention should be made of certain very regular limits between the grassy and the dense shrubby savannah. It may be concluded that there were formerly cultivated plots which have passed the fallow stage.


The geographical distribution of the crops is quite clear. They are found around the two villages and towards the west of the area.

They make up 14 percent of the whole area and thus occupy third place in terms of occupation (2).

(2) As in 1952, the same observation will be made concerning areas. The figures are probably over-high: fallows of a year have undoubtedly been confused with the year's cultivated areas.

Almost all the plots touch, at least on one side, the dense shrubby savannah, over which they have clearly prevailed. Some fields, generally the biggest, are completely surrounded by dense shrubby savannah.

The study of the forms and sizes of the cultivated areas is instructive. Most of the fields, and in particular the largest, have regular forms; the straight line dominates, and only the small plots are irregular.

This is clear proof of crop practices which are "modern" for the region: introduction of the plough and traction by animals. The already mentioned presence of livestock pens and thorn tree protection enclosures merely confirms this hypothesis.

Aerial reconnaissance eliminates any doubt on this subject, since the furrows, even after several years, remain visible as a result of low local rainfall.

Comparison of the 1952 and 1971 situations

Comparison of the two situations described above proved rather complex as soon as it was necessary to enter into detail concerning each class and its evolution (Graph no. 4).

The graph form selected here to illustrate the results of the study is fairly telling, despite a rather abstract formulation for which we apologise to the reader.

The figures carried onto graphs 4 and 5 show, in 1971:

- A fall in the areas of dense shrubby savannah.
- A rise in the total of the other classes.

Since at the outset the dense shrubby savannah was acknowledged as the original ecotype of the site, we may clearly conclude that the whole has deteriorated.

In fact, the real state of affairs is more subtle, and needs to be studied in detail. It will then be seen that while a phenomenon of deterioration of the natural vegetation has occurred (negative elements), in places one may find unarguable evidence of reconstitution of the original plant cover (positive elements).

Changes from 1952 to 1971

Taken as a whole, comparison of the areas as in 1971 and in 1952 reveals the following changes:

Dense shrubby savannah

- 23 %

Light shrubby savannah

+ 34 %

Grassy savannah

+ 37 %


+ 249 %

Both what has evolved and how it has evolved are of interest.

Graph. no. 5 and Table no. 6 give the evolution of each class, either in percentage terms or in area.

For example:

The 1952 dense shrubby savannah has evolved in such a way as to have the following composition in 1971 (1).

1,516 hectares are under crops
596 hectares are grassy savannah
2,052 hectares have become light shrubby savannah
8,723 hectares have not changed.

i.e., Deterioration of 4,164 hectares

(1) A difference of 112 hectares appears as compared with 1952. This is due to errors in plotting.

On the other hand, the following have been made into dense shrubby savannah since 1952:

176 hectares of crops
202 hectares of grassy savannah
736 hectares of light shrubby savannah
8,723 hectares unchanged

i.e. 1,114 hectares being reconstituted.

Each cartographic unit, as is shown in the tables and graphs, was processed in the same way.

To sum up, of 17,100 hectares studied, 6,303 hectares have evolved:

4,759 hectares towards deterioration;
1,544 hectares towards reconstruction.

The balance is thus negative, to the tune of 3,215 hectares.

This leads to several conclusions:

- Here is proof that the vegetation can be reconstituted; (2)
- The balance remains negative, and the causes are known: man, the plough, and livestock;
- It is the classes of vegetation most favourable to acridians, which have increased.

(2) Another piece of proof of this phenomenon: the alignment of the car track Toundé-Sagoumane has been modified; yet it is quite impossible to find on the 1971 photos the abandoned part; the other parts have become simple paths.

Thus the balance is doubly negative.


This present study is designed to present a more rigorous and especially objective approach to the evolution of vegetation on a specific piece of land.

This method is not perfect - and that is realized. In particular, the results presented here are basically "quantitative" since, in the present case, it cannot be asserted that what has been classified under "reconstitution" is exactly composed of the same elements as in 1952; only the physiognomic aspect of the plants observed is comparable. This study should have been carried out by a pluridisciplinary team, notably with a phytosociologist or a botanist.

This approach, which makes it possible to avoid the usual approximation and "impressions", can thus improve. It remains necessary to succeed in convincing those responsible that they should use it.

Carte 1 - Carte d'occupation des sols 1952-1971 - SITE DE SAGOUMANE: SITUATION EN 1952

Carte 2 - Carte d'occupation des sols 1952-1971 - SITE DE SAGOUMANE: SITUATION EN 1971

Figure 3 - Evolution de l'occupation des sols de 1952 à 1971 sur le site de Sagoumane

Figure 4 - Situations comparées de l'occupation des sols en 1952 et 1971 (en %) Site de Sagoumane

Figure 5 - Evolution de l'occupation des sols de 1952 à 1971 Site de Sagoumane en hectares

Previous Page Top of Page Next Page