SOIL LOSS BY SPLASH AND WASH DURING RAINFALL FROM 2 LOESS SOILS

Physical processes occurring during surface seal formation through a r ainstorm are well understood, but limited information is available reg arding the quantity and particle size distribution of splash and runof f at certain time intervals. In this study, we evaluated the quantity and particle size distribution of suspensions of both splash and inter rill runoff in two loess soils with different mineralogy and aggregate stability, and somewhat different particle size distribution, but sim ilar organic matter content. The soils were subjected to simulated rai nstorms of similar to 40 mm h-l and 100 mm h(-1) intensities. The amou nt of splash was about four times higher for the Saskatchewan soil (Ty pic Haploboroll) with high smectite than for the Grenada soil (Typic F ragiudalf) which is rich in Fe-oxyhydroxides. The amount of splashed m aterial and sediment load increased with increased rainfall intensity for both soils, Splash was decreased after wetting of the soil surface . The decrease in splash rate was more rapid with high rainfall intens ity. The amount of clay size particles of the splash was similar to th e original soil material. Micromorphological observations confirmed th e fluctuations in clay content with time, at the very surface. Soil ma terials splashed were much higher (10 to 20 times) than the interrill runoff losses. The latter was controlled by the rainfall intensity. Th e soil material from Saskatchewan, produced more than 11 Mg ha(-1) of interrill runoff with low rainfall intensity. High rainfall intensity produced 10 time more soil loss than low rainfall intensity. High amou nt of soil loss clearly shows that the Saskatchewan soil would benefit from erosion control measures. Both rainfall intensities removed pref erentially more clay from the Saskatchewan B horizon material. This ha s important agronomic and environmental implications for this soil. In terrill soil losses fi-um the Granada A horizon material were much les s, with no clear evidence of preferential removal of clay size particl es.