Although being temporary, the presence of tillage-induced surface roughness in the soil is an important requirement in conservation tillage systems. The reason is that surface roughness increases both surface retention and surface infiltration of water in the soil, reduces runoff velocity and volume, and traps eroded sediments, thus reducing water erosion damages. With this in mind, this study was developed with the objective of evaluating modifications in soil surface roughness by tillage and rainfall actions related to water erosion, in the absence and presence of mulch cover. The experiment was carried out in the field, at the Agriculture Experimental Station of the Federal University of Rio Grande do Sul (EEA/UFRGS), in Eldorado do Sul County, Rio Grande do Sul State, Brazil, in 1996 and 1997, using simulated rainfalls on a sand), clay loam Paleudult with 0.07 m m(-1) slope steepness. The tillage types evaluated in the study included plowing, plowing plus double-disking and no-till, all them in the absence and presence of 60% soil cover (oat residue), submitted to four simulated rainfall tests. The first test consisted of a rainfall segmented in four portions, lasting for 20, 20, 30, and 30 min, separated 30 to 40 min front each other, applied immediately after tillage. The remaining tests consisted of uninterrupted rains of 90-min duration, applied 1, 20, and 35 days after the first rain. These rainfalls were applied with the rotating-boom rainfall simulator at a constant intensity of 64.0 mm h(-1). Tillage caused greater changes in the soil surface roughness titan rainfall. Soil surface roughness was most reduced by rain action in the very first event in recently-tilled soil, in the pre-runoff period. Soil surface roughness impeded or delayed runoff ill treatments with soil Mobilization in the rainfall segments with short duration applied soon after tillage, impeding or reducing water and soil losses in that period, regardless of soil cover. In the continuous, subsequent long rains, surface roughness did not influence water loss in the studied treatments without cover, where it was high throughout the experimental period, but it did reduced water loss in the presence of cover. Water loss in no-till was high for such rains throughout the experiment. Under the same rain type, soil loss reduction as influenced by roughness was more evident in the absence of cover, whereas it was substantially obscured in its presence. Mulch of crop residue added to the soil surface did not preserve the initially high surface roughness created by tillage in the degraded soil used in the study. Nevertheless, by the end of the experiment more than half of the theoretical initial water and sediment retention capacity still remained in the microdepressions formed by roughness. The obtained data were consistent with theories and concepts used in soil erosion mechanics studies.
