TECHNICAL ADVANCE The dynamics of soil redistribution and the implications for soil organic carbon accounting in agricultural south-eastern Australia ADRIAN CHAPPELL, JONATHAN SANDERMAN, MARK THOMAS, ARTHUR READ andCHRIS LESLIE CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia Abstract Anthropogenically induced change in soil redistribution plays an important role in the soil organic carbon (SOC) budget. Uncertainty of its impact is large because of the dearth of recent soil redistribution estimates concomitant with changing land use and management practices. An Australian national survey used the artificial radionuclide caesium-137 ( 137 Cs) to estimate net (1950s–1990) soil redistribution. South-eastern Australia showed a median net soil loss of 9.7 t ha 1 yr 1 . We resurveyed the region using the same 137 Cs technique and found a median net (1990–2010) soil gain of 3.9 t ha 1 yr 1 with an interquartile range from 1.6 t ha 1 yr 1 to +10.7 t ha 1 yr 1 . Despite this varia- tion, soil erosion across the region has declined as a likely consequence of the widespread adoption of soil conserva- tion measures over the last ca 30 years. The implication of omitted soil redistribution dynamics in SOC accounting is to increase uncertainty and diminish its accuracy. Keywords: 137 Cs, Australia, caesium-137, carbon accounting, deposition, soil erosion, soil organic carbon, soil redistribution Received 9 February 2012; revised version received 9 February 2012 and accepted 10 February 2012 Introduction The potential of the land sector to contribute to mitigat- ing climate change is well recognized. The conversion of natural land to agriculture has long been recognized as capable of greatly accelerating soil erosion (Wolman, 1967) which can cause major problems for sustainable agriculture and environmental protection (Montgom- ery, 2007). Erosion in excess of soil production is widely accepted to decrease agricultural potential (Brink et al., 1977). There is concern that large magnitude, small fre- quency events (e.g., wildfires) could overwhelm green- house gas (GHG) emission reduction efforts. However, relatively small magnitude and large frequency events like soil erosion may have a similar impact on carbon accounting and GHG emissions when integrated over much longer time periods (Harper et al., 2010). The effect of soil erosion on soil organic carbon (SOC) loss is well known (e.g., Gregorich et al., 1998). Soil erosion preferentially removes the fine, nutrient-rich fraction with the potential to make significant impacts on the agricultural system without the removal of large amounts of soil. For example, assuming a conservative gravimetric SOC content of 1–5%, soil erosion from cul- tivated land of 4 t ha 1 y 1 equates to a loss of approx- imately 40–200 kg SOC ha 1 y 1 . This is comparatively large considering an average wheat crop yielding 2 t grain ha 1 y 1 would contribute approximately 800 kg C ha 1 y 1 to the soil (Skjemstad et al., 2004). The accurate and precise assessment of the magni- tude of soil redistribution (erosion and deposition) is confounded by the vagaries of the magnitude and fre- quency of events and highly variable nature of its redis- tribution in space and time (Roels, 1985). For example, soil redistribution may occur in small amounts fre- quently across large areas due to rainsplash, overland flow and tillage processes. In contrast, large magnitude wind erosion and dust emissions may remove consider- able quantities of top soil with very little visible indica- tion in the source area, but considerable downstream impact in visibility, pollution, human health, economic disruption etc. The recent (2002 and 2009) dust storms to pass through Sydney are testament to their impact on society alone (McTainsh et al., 2005; Leys et al., 2011). Soil erosion measurement and monitoring approaches, particularly in semi-arid environments, require sufficiently long (ca, 15 years) and expensive campaigns to provide representative and reliable esti- mates of soil erosion (Roels, 1985). Even with such cam- paigns the extrapolation of their results from often small experimental plots to large areas is notoriously Correspondence: Adrian Chappell, tel. + 61 2 6246 5925, fax + 61 2 6246 5965, e-mail: adrian.chappell@csiro.au © 2012 Blackwell Publishing Ltd 1 Global Change Biology (2012), doi: 10.1111/j.1365-2486.2012.02682.x