A survey of soil erosion was conducted in Australia using the fallout radioisotope caesium-137 as an indicator of topsoil redistribution. Two hundred and six sites were sampled, 100 within rotational cropping and horticultural use, 52 within uncultivated permanent pasture and forest, and 54 in rangelands. Average net soil losses were approximately equal for cultivated cropping lands and rangelands (ca. 5.5 t ha⁻¹ yr⁻¹), and just over 1 t ha⁻¹ yr⁻¹ for pasture and forest. The Mann Whitney U Test revealed that losses under cropping and rangeland conditions were significantly higher (p < 0.05) than under uncultivated pasture and forest. Soil loss was negatively correlated with mean annual rainfall and slope gradient, and positively correlated with slope length (Spearman's rank correlation). There was no correlation between rates of soil loss and a rainfall erosi-vity index. An assessment of erosional events was provided by landholders for 104 sites, with their ranking being weakly but significantly correlated with soil loss estimates (r =+0.35). Sixty percent of sites had net soil losses greater than 1 t ha⁻¹ yr⁻¹, and 74% of sites had losses of more than 0.5 t ha⁻¹ yr⁻¹. This latter rate may be regarded as a limit for a tolerable level of soil loss. These high rates of soil loss have occurred since the mid-1950s despite there being significant landholder awareness of the soil erosion hazard.
Australian Geographical Studies Vol. 42, Issue 2, p. 221-233