Soil production rates inferred from cosmogenic radionuclides, and last glacial maximum erosion rates in upland S.E. Australia

Cosmogenic radionuclides are commonly used to determine local surface lowering rates e.g., erosion rates of outcrops, and can be used to infer soil production rates by sampling saprolite and using the saprolite lowering rate as the conversion rate of saprolite to soil (Wilkinson and Humphreys, 2005). Among the various cosmic ray scaling considerations, inferring soil production rates requires an account of saprolite shielding by soil, and situations where time-invariable soil depth applies are commonly sought. However, invariable local soil depth over tens of thousands of years is unrealistic at many sites, especially where the last glacial period may have been accompanied by different rates of erosion than those observed today. Testing the constant soil depth assumption has been accomplished in two ways: morphometric analysis and slope process modelling (Heimsath et al., 1997; Heimsath et al., 2000) and, the emergence of tors from the surrounding soil (Heimsath et al., 2001). A further method involves measuring erosion rates from various parts of the catchment. For instance, in the Blue Mountains (33°S and ~ 1000 m. a.s.l. in the southeastern highlands of Australia), the erosion rate of a small catchment, determined from the 10Be concentration of stream sediment, is greater than all soil production rates from within that catchment. This casts doubt on the veracity of the apparent soil production rates since calculations of catchment erosion rates are not dependent upon a constant soil depth assumption. Apparent soil production rates may be artificially low if the modern soil depth, used in surface lowering calculations, was thinner for the bulk of the 10Be accumulation time. Considering the site was above the treeline during the LGP (Hesse et al., 2003), soil production rates are recalculated using a soil cover 30 cm thinner than that observed; this revision corresponds with the absence (and probable erosion) of saprolite below soils less than ~ 30 cm deep, and the presence of saprolite under deeper soil. Revised soil production rates overlap the catchment average; furthermore, the revised soil production function (plot of soil production rate v. soil depth) accords with models from the literature. This TCN data set adds to mounting evidence that the Blue Mountains were dramatically affected by the last glacial period. Not only was aeolian transport and sand dune formation active, but generally, soils were probably ~ 30 cm thinner on plateaux surfaces, and ~ 10% of surfaces likely experienced total soil and saprolite erosion. These phenomena were probably a result of sparse, treeless vegetation (Hesse et al., 2003) which contrasts with heath and forest that currently covers the entire plateau surface.