Independent of the runoff capacity to transport mineral and organic particles of soil, the assessment of rainfall erosive action, during its initial phase, would not be complete if the mechanisms of soil detachment and splash by raindrop impact were not appraised. Considering the lack of data on these mechanisms in soils of Ceará State, Brazil, the objectives of this work were: to compare collection methods of detached and splashed soil and to evaluate the correlations between rainfall erosivity and splash on bare and covered soil to reduce erosion in its initial phase. The study was conducted under field conditions in Fortaleza, Ceará State, Brazil, on an eutrophic Red-Yellow Podzolic during the rainy season from 3/21 to 7/9/1994, with a rain erosivity of 6,003 MJ mm (ha h)-1. In order to quantify the mass of detached and splashed soil particles, the splash cup method of Morgan (1982) and the funnel method of Sreenivas et al. (1947) were used. Confidence intervals at a level of 95 % were calculated by the t-test in paired data for the difference between means of the splash collected by the two devices. There was no statistically significant difference between both methods in the two treatments,'bare soil' as well as 'soil with a 50 % cover of napier grass stems'. Significant correlations between the EI30 erosivity parameter and the splashed soil mass for regression equations in a potency form were found.The mean of total splash collected in both methods showed a reduction of 52 % in the treatment with cover (2.31 kg m-2) when compared to that collected on bare soil (4.81 kg m-2). This result confirmed once more the effectiveness of soil cover in reducing splash erosion, which is of great importance as it is the initial stage of the erosive process.
splash erosion; soil detachment; EI30 parameter; mulch