shows some additional differentiation within the toich. 
Within the high Toich two subunits show up. The high 
Toich closest to the settlements show a dense vege 
tation cover in late Wet season only whereas the part 
further away has also some cover in mid wet season. 
Some parts of the Toich near the swamp and parts of 
the swamps (on the supervised image) appear to have 
their own seasonality and correspond with the marshy 
toich Typha-Leersia vegetation type of Yath (1984). 
The supervised classification is practically a map, 
it needs only a legend to be so. The multitemporal 
image, however, shows not such clear boundaries and 
requires some interpretation if a map is required. 
Evaluating we may say now that the multitemporal 
image is superior to both the supervised classi 
fication and the false colour composite. The multi 
temporal image might give a good land unit map on a 
scale 1:250 000. In this type of flat, fluctuating 
and flooding environment, it is also superior to 
black and white aerial photography. However, on pro 
ject scale where a land unit map of 50 000 - 100 000 
is required, only aerial photography combined with 
Landsat can give satisfactory results. 
A main limitation to produce multitemporal imagery 
is the availability of scenes of suitable season 
within one year. In fact the used imagery set of 
three dates in the only suitable one between 1972 and 
today. For other parts of the Sudan and elsewhere no 
such series might be found. 
References 
Gils, H. van & I.S. Zonneveld 1984. Vegetation and 
Rangeland Survey. ITC, Lecture note, Enschede. 
Mulder, N. 1982. Methodology of colour coding MSS 
and other data. ISP Comm. Ill, Helsinki. 
Yath, Y.A. 1984. Landscape-ecological survey of south 
Kongor, Yonglei Sudan, MSc thesis, ITC-Enschede: 
164 pp. 
Mefit-Babtie 1983. Development Studies in the Yonglei 
Canal Area. Final Report, 10 volumes. Khartoem/ 
Rome/Glasgow.