Soil science : confronting new realities in the 21st century : transactions of the 17th world congress of soil science

11 pages + poster; International audience; The aim of this study was to analyse the relationships between soil colour and the duration of soil waterlogging by phreatic or perched groundwater, in a representative hillside of the West African granito-gneissic bedrock. This duration is a decisive parameter for agricultural or forestry development, but it can vary widely over short distances on hillsides, such as those of the West African granito-gneissic bedrock. Whereas existing methods do not give ready access to this parameter, two arguments suggest that soil colour could give relatively low cost and low time-consuming indicators of the duration of soil waterlogging: i) soil color depends on the types of soil constituents present, which can depend on the duration of soil waterlogging; ii) soil color is also easily accessible and relatively stable over time. The selected catena was in central Togo lat 8°38'–8°39'N, long 1°00''–1°01''E, in a biophysical environment covering some 500,000 km in West Africa. Nineteen hydropedological stations were installed on this catena. Soil and piezometric conditions were studied on these stations for three annual cycles (1989–1992). Data collected in each station were used to calculate the mean annual rate of soil waterlogging (WLG) at 10-cm intervals from the surface down to a depth of 7 m. At the same intervals, 8 numerical colour variables were calculated from the colours of the uncemented soil phases: Munsell value, chroma, angular hue and redness rating of the principal uncemented phase; mean Munsell value, mean chroma, mean angular hue and mean redness rating. Statistical relationships were established between WLG and each of these soil colour variables. The two most significant relations were those between WLG on one hand, and mean angular hue and mean redness rating on the other hand. These two relationships provided the basis for logistic models predicting the mean annual rate of soil waterlogging. The operational limits of these models were finally discussed.