SEDIMENT YIELD PREDICTION IN TANZANIA: CASE STUDY OF DODOMA DISTRICT CATCHMENTS

Abstract

Soil erosion and sedimentation is serious and is resulting in rapid sedimentation of streams and reservoirs in Tanzania.
This makes sediment yield modeling important before undertaking water resources development projects, although it has
so far not been done in the country. At present sediment yield prediction can be done using physically based models,
empirical soil erosion and sediment yield models, and sediment rating curves. The physically based models contain
numerous parameters that are difficult to measure or estimate, making them inappropriate in developing countries like
Tanzania. The rating curves give rough estimates while the empirical sediment yield models need runoff data that is not
available in Tanzania. The revised USLE parametric (empirical) model that had been tested and found to be suitable for
soil erosion prediction was selected for sediment yield prediction study reported in this paper.
The testing of the model has shown to give acceptable results. For sediment measurements ranging from about 7
t/ha.year to 11 t/ha.year, the model predicted sediments yields ranging from about 6 t/ha.year to 12 t/ha.year, with
coefficent of determination (R 2 ) equal to about 0.53. Since soil erosion and sedimentation of water resources are serious
problems in Tanzania, sediment yield modeling using the RUSLE can be a very useful tool in the country when used.