The Federal Republic of Nigeria is the 30th largest country in the world (923,768 km²) and is 9th largest in terms of population, which amounts to about 115 million inhabitants (information from 1991).
These figures already hint at the importance of spaceborne information for the sustainable development of Nigeria.
The country's major resource lies in the exploitation of offshore oil fields. Due to its fast growing population, the management of agriculture as well as the development of forest resources are a major challenge for the national and regional administrations. The availability of accurate and appropriate topographic maps can be considered absolutely essential for the country. As far as expertise in remote sensing is concerned, Universities (e.g. Lagos, Ile-Ife) have been able to increase their capabilities through cooperative programmes with different countries.
The Regional Centre for Aerospace Surveys in Ile- Ife has been successfully employed as the motor behind this development of earth - observation activities. In addition, Nigeria has recently created a national remote-sensing centre in its new capital, which shows the increasing involvement of the Nigerian government. Besides national administrations, private companies (national and foreign) are very much interested in information about the development status of a particular region.
The availability of a land cover map and its introduction into a GIS (Geographic Information System) can be considered as one basis for the sustainable planning and management of the resources of a particular country or region. The best scale for such a land-cover map depends on the particular application and the availability of data to be interpreted. For information needs at a national level, a scale of 1:500,000 should be sufficient, taking the size of Nigeria into account. The class requirements (nomenclature) depend first of all on the specific needs of a particular administration. However, a number of basic land cover classes should be included in all maps (e.g. rivers, lakes, roads, agriculture, forests, urban areas). In this respect, the topographic and land- cover maps carry very similar information (except for the height information of topographic data sets).
Regional land cover maps should give a more detailed view and a scale of 1:50,000 to 1:100,000 is therefore more appropriate. Again, the number of classes needed depends on the goals of the work (e.g. planning a road) but data used to archive the class resolution has to be investigated very carefully. Not all of the desired classes can be detected from space.
Therefore, a certain compromise between the requested resolution and the detectability of a particular object sometimes has to be made.
The ERS SAR can be used to produce land-cover maps at national level and with a scale of 1:500,000 without any problem while taking the above described broad classes into account. Image 6 is a full ERS-SAR image (100 x 100 km) taken over the coastal zone in the south of Nigeria. For visual interpretation purposes, it has been filtered with a 10 x 10 pixel moving box average filter.
Image 6: ERS-SAR image over the south of Nigeria of August 14,
1994
The most important and most visible features in the image are certainly elements of the hydrologic network such as rivers, streams, lakes and lagoons, which can be detected and mapped very easily. The coastline is also clearly visible.
Urban areas can be distinguished as white spots, like the town of Benin City in the upper right corner. Smaller villages and towns can be seen by increasing the screen resolution (image 6 has been reduced by a factor 12, in order to display the full data set on the screen).
In the upper centre a square surface with a different grey level can be seen. Looking at a subset, one can identify this feature as a forest cutting (see image 7).
Image 7: ERS subset with forest cutting visible on the right
side of the image
Image 8 shows the same feature in a multitemporal data set taken from different orbits. Although the river appears in black (a sign of successful superposition), red and blue lines can be observed and interpreted as a sign of a bad superposition. This low georeferencing quality is caused by the mountainous terrain conditions affecting the upper part of the image. In this case, georeferencing with the help of a DTM (Digital Terrain Model) would give a better result.
Image 8: Terrain disturbed multitemporal image
However, even with the distortion, the forests and forest cuts are clearly visible on the image. As these surfaces were also observed in other images, ERS-SAR can be seen as an appropriate tool for the management of forest concessions (clear cuttings) for the whole of Nigeria. Improved image interpretation capability through the multitemporal approach can already be seen in image 8. This technique is even more valuable for agriculture as features change more quickly than in forestry or cartography.
Image 9 is an example of ERS-SAR data's potential for monitoring rice. Areas in dark red are surfaces under rice. Due to the rice-cycle, the areas are flooded and therefore easily detectable. The transverse line (upper left to lower right) which divides the image into a more bluish and a more reddish area is probably the boundary between cultivated (lower part) and forested land. As the individual fields are smaller than the resolution of the ERS-SAR data, the boundaries between them cannot be detected.
Image 9: Agriculture areas with rice fields (dark reddish
spots)
Image 10 shows the total overlapping area of two images from a descending and an ascending orbit. The boundary between the forested and agricultural land can again be clearly identified, whereas the reddish coloured surface corresponds to the cultivated area. White spots distributed over the whole image are small villages. The red area in the lower left part is the ocean. A clear difference between open water and the coastal lagoon is visible and could be mapped.
Image 10: Multitemporal data set of images acquired on August
2 and 14, 1994
Besides the ERS-SAR's potential for mapping the hydrological network of a country or region, flooding phenomena can also be monitored with its data on a regular basis, as the next example shows. Whereas the first image (11) was acquired before the event on August 2, the second subset (image 12) was apparently taken during the event or shortly afterwards (August 14); the black colour shows that the area is under water.
The multitemporal data (image 13) s et gives a good view of the original river bed and its extension during the flooding event.
Image 11: Area before flooding
Image 12: Area during or shortly after flooding
Image 13: Synthesis by multitemporal approach
SP-1199