Effectiveness of soil conservation measures in two contrasting landscape units of South Eastern Tanzania

RUSLE, the revised universal soil loss equation, is widely used for estimating potential soil erosion by water. Field measured model factors are however scarce for tropical regions. We derived RUSLE factors for several soil conservation measures based on three seasons of field plot measurements (2008-2010), in two contrasting landscape units of South Eastern Tanzania, the country's most important cashew growing area. Whereas the derived factors are useful for applying RUSLE in similar environments, this study points to the importance of understanding differences between soil types and landscape units when assessing potential soil erosion. On the Makonde plateau, rainfall erosivity was higher than on the inland plains (7,130 vs 5,783 MJ mm ha-(1)h(-1) year(-1)). The soil erodility K factor was also higher (0.014 t h MJ(-1) mm(-1)) on the sandy Cutanic Acrisols of the Makonde plateau than on the clayey Acric Ferralsols (0.006 t h MJ(-1) mm(-1)) of the inland plains. Likewise, soil loss on the Makonde plateau was much higher than on the inland plain (e.g. for maize 33-127 t ha(-1) season(-1) vs 3-10 t ha(-1) season(-1)). The differences between the C factor for maize, and for maize with crop residues, as well as the differences between the P factors for lemon grass strips and ridges and furrows indicate that although soils of the Makonde plateau are more susceptible to soil erosion, these soils are also more responsive to soil conservation measures compared to soils of the inland plains. Farmers' local technique of making ridges and furrows is particularly effective. Furthermore, the C factors for cashew groves are one order of magnitude lower (0.08-0.09) than for maize, or for maize with crop residues (0.2-0.7), but are still much higher than for bush fallow (0.001). Deforestation for agriculture hence bears the risk of increasing soil erosion rates; however, this risk can be minimised by growing cashew trees.