Charles CAZABAT *
(*) Charles Cazabat: Engineer, Institut Géographique National, 136 bis. rue de Grenelle, 75007 Paris.
Data on the Sahel
Proposals for a methodology
The author, who is a Survey Engineer at the Remote Sensing Department of the National Geographical Institute in Paris and co-investigator of the ERTS project, presents the different recording techniques using satellite, aeroplane and field observations. He analyses the advantage of each method of data collection and proposes scales (for use in automatic or traditional methods of cartographic presentation appropriate for the requirements of the data and area to be covered).
He then proposes a methodology involving three stages, which will make it possible to arrive at ground level operations:1) A general description of the Sahel, at l: 500,000, obtained through the checking of results by intersection from satellite data and data already available in quantity on the countries involved.
2) A description of the rangeland, at 1: 200,000, obtained through normal procedures of photo-interpretation, which should produce a map showing the suitability of the land for pasture, and giving the current potential of the regions.
3) A study of the evolution of rangeland and the monitoring of rangeland based on detailed observations made in limited areas and transects chosen for their "sensitivity," which would require the setting up of centres for the annual forecasting and management of the rangeland.
In conclusion, the author shows that the problem is essentially a political one; and that it will require the training of many specialists and above all, the drawing up of an overall physical planning programme.
Pastoral grazing migrations are a basic feature of African livestock activity, and their routes lie mainly across the Sahelian region.
The existing structures are due to an adaptation over centuries to natural and quasi-permanent drought conditions; and any solution of these problems that would entail radical modifications of the current structures would, without doubt, pose many more problems than it would resolve.
Actions to be undertaken must take account of this aspect, and research must be exclusively directed towards the improvement, at all levels, of the precarious existing conditions, i.e.:
a) improvement of the data on the natural conditions of the region,
b) improvement of the grazing lands and their use,
c) improvement of the livestock,
d) improvement of the human conditions linked with the migrations, and
e) improvement of the commercial circuits.
This paper will be restricted to the improvement of the data on the natural conditions, and will attempt to discuss the methods currently available for improving data collection in the Sahelian region for improving the definition of pasturelands, and for improving the mapping of them; after which a general methodology based on a practical cartography will be proposed as a preliminary to ground operations.
Excluding maps already out-of-date at the 1: 200,000 or 1: 250,000 scales, one can state that with the exception of a few publications at even smaller scales, there are no really accurate documents describing and giving data on the Sahelian region.
Current aerial photography and remote sensing techniques would, without doubt, enable one to reduce this omission quite quickly; and if they did not produce a complete account of the region, they would at least give a uniform description of the African pasturelands.
A considerable number of recording processes, vectors, and sensors are currently being used for remote sensing purposes, but they can be grouped into 3 main categories that are really complementary:
a) imagery from satellites,
b) recordings made on board aircraft,
c) field observations.
The Earth Resources Technology Satellite ERTS 2, which is likely to be launched in March 1975, is the most appropriate system for analysing major features and at the same time obtaining a global synthesis of groups of features.
Having a technology similar to that of ERTS 1*, the ERTS 2 satellite will have the same sensors, and the recordings will be made using two different processes:
(*) Ch. Cazabat, P. Demathieu, A. Dupuis, F. Verger. The FRALIT Programme: Remote sensing of the French Atlantic Littoral by the ERTS 1 Satellite. (In French), I.G.N. Information Bulletin No. 19, September 1972.
a) Return Beam Vidicon (R.B.V.) system, similar to a television camera, in three bands of the visible spectrum:Channel 1: wavelengths between 475 and 575 nanometres;
Channel 2: wavelengths between 575 and 680 nanometres;
Channel 3: wavelengths between 680 and 810 nanometres.
b) Multi-Spectral Scanner (M.S.S.) system, working in four bands, one of which covers a large part of the near infrared region:Channel 4: wavelengths between 500 and 600 nanometres;
Channel 5: wavelengths between 600 and 700 nanometres;
Channel 6: wavelengths between 700 and 800 nanometres;
Channel 7: wavelengths between 800 and 1,000 nanometres.
These images are an important advance because:
1) Large areas are surveyed simultaneously
The apparent focal lengths of the sensors and the 900 km orbit around the earth enable large areas to be surveyed; more than 34,000 sq km (185 x 185 km) are recorded homogeneously- in one block on a single image.
2) The data are repetitive
Every 18 days the satellite passes exactly over a given point; consequently it is possible to follow accurately the detailed evolution of pasturelands, surface hydrology, and drought conditions in a particular region.
3) The information is selective
Each of the various Channels records different data **:
(**) Ch. Cazabat: The importance of the equidensity method and its application to the ERTS-FRALIT Programme. (In French), Bulletin de l'Association de géographes français, No. 411-12, Nov.-Dec. 1973.
Channels 1 and 4 (blue-green) integrate data relating to water vapour, nature of the ground, transparency of water, and transportation of sediments.
Channels 2 and 5 (yellow-orange) record information similar to that obtained with normal panchromatic emulsions, and are consequently very useful for studies of the nature of the ground and crops, natural vegetation, and the tree and shrub density in wooded regions.
Channels 3 and 6 (red and very near infrared) record data similar to those obtained by Channels 1, 2, 4 and 5.
Channel 7 (near infrared) enables one to differentiate between the various effects of the drought conditions on the vegetation and to detect surface water.
The electromagnetic recording of the data enables one to process them in various manners:
1) Digital cartography
Where the original tapes are digitized and processed using computer programmes, plotting tables, or printouts enabling one to analyze various aspects of a given theme.
Computer mapping requires a fairly long running in time, but it does allow one to set up cartographic groups responsible for large areas at scales between 1: 500,000 and 1: 100,000. Wishing to publish at larger scales is futile because of the low resolution of the sensors and the poor contrasts of the Sahelian region.
2) Conventional cartography
The photo-interpretation of images which have been decoded and photographically processed at scales between 1: 1,000,000 and 1: 200,000 can be carried out in the conventional way.
Photographic mosaics can be used as base maps and allow one to make detailed interpretations for each channel; however, since the information is spread over 7 channels it is necessary to produce interpretation sketches for each theme (geomorphology, pedology, potential use of the land for pasturage, etc.).
One must remember that the resolution is not good and that numerous details of the planimetry are invisible, such as the tracks and isolated dwellings.
3) Improved mapping
Using both conventional photo-interpretation techniques and those of equi-densities, either by computer processing or chemical treatment **, together with the fine grain emulsions enables one to make selective masks and direct mapping at the 1: 1,000,000 to 1: 500,000 or even 1: 200,000 scales without real difficulty.
(**) Ch. Cazabat: The importance of the equidensity method and its application to the ERTS-FRALIT Programme. (In French), Bulletin de l'Association de géographes français, No. 411-12, Nov.-Dec. 1973. scanners, and some, such as the Daedalus, work equally well in the visible near infrared and thermal regions.
Images obtained using aircraft
Aircraft enable one to record data from large areas at scales between 1: 2,000 and 1: 200,000.
The sensors used are of two types: electromagnetic and photographic.
Based on the same principles as the satellite equipment, the electromagnetic sensors are multi-spectral scanners, and some, such as the Daedalus, work equally well in the visible near infrared and thermal regions.
For example, it is possible to record on the following channels:
Channel 1: 380420 nanometres visible spectrum
Channel 2: 420-450 nanometres visible spectrum
Channel 3: 450-500 nanometres visible spectrum
Channel 4: 500-550 nanometres visible spectrum
Channel 5: 550-600 nanometres visible spectrum
Channel 6: 600-650 nanometres visible spectrum
Channel 7: 650-700 nanometres visible spectrum
Channel 8: 700-800 nanometres visible spectrum
Channel 9: 800-900 nanometres visible spectrum
Channel 10: 900-1,100 nanometres, near infrared
Channel 11: 3 to 5 micrometres thermal
Channel 12: 8 to 14 micrometres infrared
For surveying pasturelands, Channels 1 to 4 are of little interest because of the very strong reflectance of the sands and also the atmospheric problems; on the other hand, Channels 5 to 8, covering the region from green to red, are very useful, like Channels 9 and 10 in the infrared, for showing vegetation and the effects of drought. Channels 11 and 12, the thermal ones, provide detailed information on the temperatures of water, ground, and vegetation, and also on evapotranspiration, which is related to the ability of the pasturelands to retain their stored water.
Taking into account the small width covered by the recording, e.g. 6 km on a flight at 3,800 m altitude for an apparent scale of 1: 100,000 and a ground resolution of 10 m, one sees that this process can only be used to cover small zones or to survey pro files. It cannot be used for basic cartography, but it is a fundamental tool for working in areas of agrostologic studies on the ground.
The emulsions normally used - panchromatic, infrared, colour, false-colour - and their applications are known by everyone. One must not be too tempted by the infrared or false-colour emulsions for use at small scales (1: 100,000 or 1: 50,000) because of the very strong reflectance of the Sahelian ground and the lean pasturelands. However, they are useful for areas of dense pastures and at scales greater than 1: 20,000; at this latter scale comparisons with ground studies enable one to carry out original research programmes *.
(*) Peyre de Fabrègues B., Rossetti C., 1971. Natural Sahelian pasturelands of Sud Tamesna (Rep. of the Niger). Evolution of the pasturelands. Mapping of pastoral potential and evaluation of fodder production by aerial photography. (In French), I.E.M.V.T. Geotechnip. Et. Agrostologique No. 32, 1971.
Aerial photography can give rise to photomaps that can be immediately used in the form of mosaics.
The scale of 1: 100,000 for the making of the models and publication at the 1: 200,000-scale are suitable for geomorphological, morpho-pedological, potential use of soil purposes, etc. They provide a solid basis for the production of maps of the existing pasturelands.
Photographic documents at scales of the order of 1: 200,000, used in conjunction with detailed ground studies, also enable one to follow the evolution of the same pasturelands over the years (the encroachment of the desert, etc.).
The belief that one can undertake photo-interpretation studies solely with satellite or aerial photo graphs and avoid simultaneous ground surveys belongs not to the domain of remote sensing but to that of psychic sensing; one cannot overstress the fact that serious research must be based on a very detailed and extensive knowledge of local ground conditions. This information is available at two levels:
1) General information of an interdisciplinary nature concerning the entire pastoral zone, enabling one, using the photo-interpretation and ground truth techniques, to have an overall picture of the surface conditions valid for many years to come.
2) Special information, detailed in nature, of the evolution of the pasturelands. This latter type of information can only be obtained by systematic studies of a restricted area over several years based on stock forms; these operations require considerable logistic support, but they are very necessary in every country. Attempts to avoid such support services will inevitably lead to failures.
Taking into account the above discussions, a methodology can now be proposed; its aim would be to closely associate the satellite and aircraft remote sensing techniques with the research programmes of the agrostologic stations on the ground. It would have three main aims:
1) A general description of the Sahelian region: Without wishing to enter into a complete description of the Sahelian natural resources, which would be a research project in itself, a preliminary definition of the geomorphological and pedological criteria of the zone would be necessary.
The satellite imagery, enlarged to 1: 500,000, for example, would provide a good basis for further developments; and afterwards 1: 100,000-scale mosaics based on stereoscopic aerial photographs would form a suitable basis for detailed studies.
The final mapping could be published at the 1: 200,000-scale, which would give rise to between 120 and 140 sheets for the entire Sahelian zone. It would show details of the hydrography, roads, tracks, dwellings, water holes, administrative limits and place-names.
Transparent overlays or other maps would show geomorphological, hydrological and sedimentological data.
Such a project would entail about 5 years of work.
2) A description of the pasturelands
Starting from the above documents, a special study could be done at the same time of the potential of the ground for use as pastureland, with a description of the various types encountered as a function of the nature of the soils.
It could be based on the previous recording complemented by a series of profiles at variable distances and by various procedures, depending on the landscape unit groups already defined. The ground operations and the determination of the interpretation keys would enable one to extrapolate the ground observations over the whole region.
Teams of geomorphologists, pedologists and agrostologists would be necessary. They could produce more detailed interpretations if their studies were based on special stations.
3) The evolution of the pasturelands
The potential having been defined, a third study phase seems obligatory, i.e. the evolution of the pasturelands: these are closely conditioned not only by the annual phenomena of rains and droughts, but also by effects of grazing, tracks, and resting-places.
A group concerned with planning and annual organization of tracks should be set up at the same time that the agrostologic studies of their evolution, the studies of potential cattle density, and the bromatological improvements of livestock are carried out.
These studies should be carried out jointly in the different countries of the Sahelian zone; they would use remote sensing techniques not only for study of limited zones but also for the imagery of large areas. For these annual data collections, the ERTS 2 satellite could be of further use, particularly if ground truth was available from rapid surveys using helicopters or from aircraft flying different transects whose "sensibility" has been established during the first study.
The methodology closely associates remote sensing procedures with ground operations; it also requires the effective participation of the agricultural and stock farming services of different African countries and various research institutes.
The carrying out of descriptive and evolutionary projects only has value if they lead to effective actions and if they influence the routes and the resting periods of the migrations across the pasturelands.
Consequently there must be very close coordination between the research institutes and the local authorities; at this level the problem is basically a political one and requires, if it is to be resolved, not only a choice and a will, but also a programme with a timetable and the training of specialists in each country. It is reasonable to believe that a period between 5 and 10 years will enable one to succeed in arranging the required close cooperation between governments and national and international organizations.
It is a hope that I express in the name of the French organizations and in particular the Institut Géographique National, various members of which are ready to collaborate in such a project.
1. CAZABAT, Ch, DEMATHIEU, P., DUPUIS, A. et VERGER, F. - Le Programme FRALIT Télédétection par le satellite ERTS 1 du littoral océanique de la France. (Bulletin Information IGN, n° 19) Septembre 1972.
2. CABAZAT, Ch. - L'intérêt de la méthode des équidensités, application au programme ERTS-FRALIT. (Bulletin de l'association de géographes Français n° 411412). Nov.-Déc. 1973.
3. PEYRE DE FABREGUES; B. et ROSSETTI, C. - Pâturages naturels sahéliens du Sud Tamesna (Rép. du NIGER. Evolution des pâturages. Cartographie du potentiel pastoral et évaluation de la production fourragère par photographie aérienne I.E.M.V.T. Geotechniq (Et. Agrostologique n° 32), 1971.