The Earth's surface erodes by processes that occur over different spatial and temporal scales. Both continuous, low-magnitude processes as well as infrequent, high-magnitude events drive erosion of hilly soil-mantled landscapes. To determine the potential variability of erosion rates we applied three independent, field-based methods to a well-studied catchment in the Marin Headlands of northern California. We present short-term, basin-wide erosion rates determined by measuring pond sediment volume (40 years) and measured activities of the fallout nuclides 137 Cs and 210 Pb (40-50 years) for comparison with long-term (> > > > >10 ka) rates previously determined from in situ-produced cosmogenic 10 Be and 26 Al analyses. In addition to determining basin-averaged rates, 137 Cs and 210 Pb enable us to calculate point-specific erosion rates and use these rates to infer dominant erosion processes across the landscape. When examined in the context of established geomorphic transport laws, the correlations between point rates of soil loss from 137 Cs and 210 Pb inventories and landscape morphometry (i.e. topographic curvature and upslope drainage area) demonstrate that slope-driven processes dominate on convex areas while overland flow processes dominate in concave hollows and channels. We show a good agreement in erosion rates determined by three independent methods: equivalent denudation rates of 143 ± ± ± ± ± 41 m Ma − − − − −1 from pond sediment volume, 136 ± ± ± ± ± 36 m Ma − − − − −1 from the combination of 137 Cs and 210 Pb, and 102 ± ± ± ± ± 25 m Ma − − − − −1 from 10 Be and 26 Al. Such agreement suggests that erosion of this landscape is not dominated by extreme events; rather, the rates and processes observed today are indicative of those operating for at least the past 10 000 years.well-studied, soil-mantled upland landscape where long-term erosion rates were previously quantified using in situproduced cosmogenic nuclides (Heimsath et al., 1997(Heimsath et al., , 1999. We focus on quantifying short-term rates using two independent methods and then compare the results with the long-term rates. Specifically, we use short-lived, falloutderived nuclides as well as measurements of sediment stored in a stock pond.Anthropogenic 137 Cs (t 1/2 = 30·1 years) and naturally occurring excess 210 Pb (half-life, t 1/2 = 22·3 years; 'excess' is the amount of 210 Pb above the level supported by soil 222 Rn) have gained wide use in recent years as tracers of soil movement (e.g.