Cosmogenic nuclides in stream sediment are widely used to quantify catchment-average erosion rates. A key assumption is that sampled sediment is representative of erosion from the entire catchment. Here we show that the common practice of collecting a narrow range of sizes—typically sand—may not yield a representative sample when the grain size distribution of sediment produced on slopes is spatially variable. A grain size bias arises when some parts of the catchment produce sand more readily than others. To identify catchments that are prone to this bias, we used a forward model of sediment mixing and erosion to explore the effects of catchment relief and area across a range of altitudinal gradients in sediment size and erosion rate. We found that the bias increases with increasing relief, because higher-relief catchments have a larger fraction of high elevations that are underrepresented in the sampled sand when grain size increases with altitude. The bias also increases with catchment area, because sediment size reduction during transport causes an underrepresentation of more distal, higher elevations within the catchment. Our analysis indicates that grain size bias may be significant at many sites where cosmogenic nuclides have been used to quantify catchment-average erosion rates. We discuss how to quantify and account for the bias using cosmogenic nuclides and detrital thermochronometry in multiple sediment sizes.
Lukens, C. E., Riebe, C. S., Sklar, L. S., Shuster D. L. (2016): Grain size bias in cosmogenic nuclide studies of stream sediment in steep terrain . Journal of Geophysical Research: Earth Surface. 121 (5), DOI: 10.1002/2016JF003859. DOI: 10.1002/2016JF003859