Submitted to: Journal of Geophysical Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/29/2005
Publication Date: 12/7/2005
Citation: Nearing, M.A., Kimoto, A., Nichols, M.H., Ritchie, J.C. 2005. Spatial patterns of soil erosion and deposition in two small, semiarid watersheds J. Geophys. Res., 110, F04020, doi:10.1029/2005JF000290.
Interpretive Summary: This study was undertaken to use radioactive Cesium in soils to measure the distribution and rates of soil erosion in two small semi-arid watersheds located in southeastern Arizona. The radioactive Cesium that we measured was deposited in soils across the entire world as a result of atmospheric atomic bomb testing that was conducted by various nations in the period largely around the early 1960s. After that time bomb testing was largely restricted to underground tests, specifically to reduce the release of such radioactive material into the atmosphere. For a long time scientists have been measuring the amount of sediment that leaves a watershed. This is important to know because sediment caused by soil erosion is a major contributor to non-point source pollution of America's rivers, streams, and reservoirs. However, the amount of sediment that a watershed generates to river and stream systems does not tell the whole story about soil erosion. What scientists are generally not able to measure on a routine basis are the hillslope erosion rates and the spatial distributions of erosion in watersheds. Yet such knowledge is important for two reasons. First, in order to treat watersheds so as to reduce sediment generated, we must know in general where the sediment is coming from. Secondly, since we use computer models to help us assess erosion in watersheds, and the positive impacts of the application of conservation efforts on the landscape, we need data to help us know whether the models are working correctly. Our results indicate that erosion rates in these watersheds were actually much greater on average than we expected for rangelands. Just as interesting was the fact that the measurement of sediment leaving a watersheds told us very little about the amounts of erosion that were taking place on hillslopes within the watersheds. In this case the amount of sediment measured in the traditional manner at the watershed outlets were extremely different from watershed to watershed, while erosion rates on the hillslopes inside the two watersheds were not that much different at all. This was due to the fact that most of the sediment generated in one watershed was deposited before it left the outlet, while nearly all of the sediment in the other watershed left the watershed outlet. This means that we must be very careful when we interpret sediment yield rates from watersheds. This study has significant implications for improving our ability to manage the soil and water resources of this nation by improving our knowledge of erosion rates in rangelands of southern Arizona and providing spatial data needed to test and improve the tools we use for conservation planning.
Technical Abstract: In semiarid ecosystems, soil erosion is influenced by the dominant vegetation of the ecosystem, soils, and landscape morphology. Rates and spatial patterns of soil redistribution were measured using the 137Cs (137Cesium) technique in a 3.7 ha, shrub dominated watershed and a 1.9 ha, grass dominated watershed in the Walnut Gulch Experimental Watershed in southeastern Arizona, USA. The calculated soil redistribution rates ranged from -9.8 to +5.7 t ha-1 year-1 and -7.9 to +10.1 t ha-1 year-1 for the shrub and grass watersheds, respectively. The mean soil loss rates in areas of net loss were -5.9 and -3.2 t ha-1 year-1, and the calculated net erosion rates for the entire watershed, including depositional areas, were -4.27 and nearly zero t ha-1 year-1 for the shrub and grass watersheds, respectively. The results indicated that both soil loss and sediment yield from the shrub watershed over the past 40 years were greater than that from the grass watershed. However, erosion rates in the grass watershed were relatively great considering the fact that sediment yields from that watershed were very small, which indicates that sediment yield from a watershed outlet can be a poor indicator of hillslope erosion rates within the watershed. The soil loss difference between watersheds was probably attributable to the differences in vegetation, while the sediment yield difference was largely controlled by the existence or lack of incised channels. The calculated net soil redistribution rates were dependent on the percent of rock fragments in the top 25 cm regardless of the different vegetation ecosystems.