Soil erosion on hillslopes occurs by processes of soil splash from raindrop impacts and sediment entrainment by surface water flows. This study investigates the process of soil erosion by surface water flow on a stony soil in a semiarid environment. A field experimental method was developed whereby erosion by concentrated flow could be measured in predefined flow areas without disturbing the soil surface. The method allowed for measurements in this study of flow erosion at a much wider range of slopes (2Á6 to 30Á1 per cent) and unit discharge rates (0Á0007 to 0Á007 m 2 s À1 ) than have been previously feasible. Flow velocities were correlated to discharge and hydraulic radius, but not to slope. The lack of correlation between velocity and slope might have been due to the greater rock cover on the steeper slopes which caused the surface to be hydraulically rougher and thus counteract the expected effect of slope on flow velocity. The detachment data illustrated limitations in applying a linear hydraulic shear stress model over the entire range of the data collected. Flow detachment rates were better correlated to a power function of either shear stress (r 2 = 0Á51) or stream power (r 2 = 0Á59).

[1] This study reports sediment yields from seven small (0.18-5.42 ha) watersheds in Southern Arizona measured from 1995 to 2005. Sediment concentrations and total event sediment yields were related to storm-runoff characteristics, and statistical relationships were developed to estimate sediment yields for events with missing data. Precipitation ranged from 263 to 298 mm yr À1 , runoff ranged from 8.2 to 26.4 mm yr À1, and sediment yields ranged from 0.07 to 5.7 t ha À1 yr À1 , with an areal average of 2.2 t ha À1 yr À1 . For six of the seven watersheds, between 6 and 10 events produced 50% of the total sediment yields over the 11-year period. On the seventh watershed, two storms produced 66% of the sediment because of differences in the geomorphology and vegetation characteristics of that area. Differences between sediment yields from all watersheds were attributable to instrumentation, watershed morphology, degree of channel incision, and vegetation.

No abstract

The Water Erosion Prediction Project (WEPP) is a new water erosion prediction technology being developed by the USDA-Agricultural Research Service to replace the Universal Soil Loss Equation. Rangeland field experiments were designed to parameterixe the WEPP rangeland erosion model. Included in the field experiments were plot treatments designed to separate direct from indirect effects of vegetation canopy on runoff and soil erosion. Nine rangeland sites from a wide range of soil and vegetation types were evaluated using rainfall simulation techniques. Natural versus clipped treatment surface characteristics and runoff and erosion responses were compared using regression analyses. These analyses showed that there were no significant differences between natural and clipped plot surface characteristics, runoff ratios, final infiltration rates, or initial rainfall abstractions. Erosion rates were different between treatments with the clipped plots having slightly less erosion than the natural plots. Results indicated that, under the rainfall conditions simulated, canopy cover was not directly contributing to initial abstractions through rainfall interception loss or significantly affecting runoff or erosion.

Critical sh(川r stress and critical flow rates for initiation of rillin日 .r Hydrol ' 142: 251-271 This study was conducted to identify critical shear stress and critical flow rates re叶uired to initiate rilling on 比lected sites. The data used in this investigation were collected from soils located throughout the US 八 认 h e re crop residues had oeen removed 、and moldooard plowing 且nιI disking had occurred. RunolLlIld 出 ， i! l(1sS measurements were made on sites where simulated rainfall was applied to pre! 川med rills. Multiple rcgrcs川on analyses were used to relate critical shear stress values and critical Il o 认 仁l t e s to selected 刊oil properties , The s(丁il-oased regression equations were found to provide reli 川、 Ie estimates. Informati 川1 identilied in this study will improve our aoility to understand and pr仆perly model upl 川ld runolr and erosion processes Several equations for estimating bed load sediment transport have been