Selected Interpretation Results:

CENTRAL AFRICAN REPUBLIC

The Central African Republic has a surface area of 622,984 km² and about 2,9 million inhabitants. Its major income comes from the exploitation of minerals. Figure 10 is a geographic overview map of the country.

Figure 10: Geographic map of the Central African Republic

The mapping of geological features related to the potential exploitation of minerals can be considered as very important for the economies of Central African countries that are not industrialized.

Radar data have been used for the mapping of geological features. These data were acquired during airborne campaigns and interpreted for detailed geological surveys aimed at the preparation of mineral exploitation. The availability of large scale (compared to aerial photography) ERS-SAR data is very much complementary, as a first survey can be made with satellite data in order to identify areas of major interest, which can then be mapped from aircraft in more detail. This complementarity of methods can save a lot of money.

An ERS-SAR image of the Carnot region acquired on August 8, 1994 during a descending orbit was analysed in detail by the LGGST (Laboratoire de géologie-géomorphologie structurale et télédétection, Université Pierre et Marie Curie, Paris) in order to evaluate and demonstrate the general interest of using radar data for geological purposes in Central Africa.

For this analysis, two geological maps provided by Barbet (1939) and Gérard (1950), being the most recent products, were used as reference data. Image 34 clearly shows three major fractions: NS, NE-SW and NW-SE. It is difficult to define the chronological relation between the different directions, but one could imagine that the NS structures were disturbed by the NE-SW fracture.

Image 34: Full ERS-SAR image of the Carnot region (Central African Republic) acquired on August 8, 1994 principal structural direction are indicated with white arrows

In images 35 and 36, a distinction between the gneisic bedrock formation (Gn) and the above lying standstone formation (Gr) can be made. This sandstone (Carnot) has a smooth texture, disturbed by a network of thalwegs and drainages all located within 2 to 3 kms.

Image 35 : ERS-SAR subset

Image 36: interpreted geological map - Gn: Gneiss, Gr: Sand stone

The gneisic formation, characterized by higher retrodiffusion values (visible in a light grey), shows a rougher texture and a denser hydrological network (1 km). Thus, one can locate the boundary between the two formations with an accuracy of about 1 km. Within the bedrock formation, a number of lithological units can be distinguished, but boundaries are very difficult to find. The two different types of gneis (indicated in the map) can also be separated.

Images 37 and 38 show the difference between the underlying quartz schist (Qs) and the covering sandstone formation (Gr). The underlying formation has a very specific texture of turned lines due to quartz banks that form small slopes oriented towards the West. These different lithological units can be mapped by taking into account mainly the texture and the hydrological network.

Image 37: ERS-SAR subset

Image 38: interpreted geological map Gr: Sand stone, Qs: quartzid

Within images 39 and 40 a large number of folds (east-west oriented) can be observed in the bedrock formation. These folds, not indicated at all on the reference maps, should be studied in more detail in the field.

Image 39: ERS-SAR subset

Image 40: interpreted geological map A: Amphibolites, Rg: rocks

Even this short analysis of the ERS-SAR imagery over the Central African Republic shows the high potential of radar data for geological mapping, providing new information not shown on existing maps. Based on the radiometric and geometric characteristics of the ERS-SAR (sensitive to surface roughness and relief variations, type of hydrological network, etc.), a number of new geological and geomorphological units can be identified. The different rock types can be distinguished, as well as the different fractions and folds. Thus, ERS-SAR images provide new information that complements the available maps. Although the general texture observed in the image can also be found in the map, details vary a great deal. Specifically, the extension and boundaries of the different features seem to be identified much less accurately on the map. These differences should, however, be validated by a field trip.

In a region mainly covered by dense vegetation (like the region studied), the identification of lithological units is possible by mapping the texture expressed by the hydrological net work. This mapping can easily be performed by a visual interpretation of a georeferenced image. If a more sophisticated approach is to be applied, one can use programmes extracting textures in an automatic way, followed by a segmentation that recognizes homogeneous forms.

In conclusion, it can be underlined that geological applications in Central Africa can benefit greatly from the use of ERS-SAR data. The above images and their interpretations show that spaceborne radar data is the only practical way of complementing existing geological maps, as the almost permanent cloud cover does not allow the use of optical sensors. In addition, large areas can be covered at relatively low cost by exploiting satellite images, which favours their use in Third World countries.