Evaluation of the 137Cs techniques for estimating wind erosion losses for some sandy Western Australian soils

The utility of the caesium-137 (137Cs) technique, for estimating the effects of wind erosion, was evaluated on the soils of a semi-arid agricultural area near Jerramungup, Western Australia. The past incidence of wind erosion was estimated from field observations of soil profile morphology and an existing remote sensing study. Erosion was limited to sandy surfaced soils (0-4% clay), with a highly significant difference (P < 0.0001) in 137Cs values between eroded and non-eroded sandy soils, with mean values of 243 ± 17 and 386 ± 13 Bq m-2 respectively. Non-eroded soils, with larger clay contents, had a mean 137Cs scontent of 421 ± 26 Bq m-2, however, due to considerable variation between replicate samples, this value was not significantly different from that of the non-eroded sands. Hence, although the technique discriminates between eroded and non-eroded areas, the large variation in 137Cs values means that from 27 to 96 replicate samples are required to provide statistically valid estimates of 137Cs loss. The occurrence of around 18% of the total 137Cs between 10 and 20 cm depth in these soils, despite cultivation being confined to the surface 9 cm, suggests that leaching of 137Cs occurs in the sandy soils, although there was no relationship between clay content and 137Cs value for either eroded or non-eroded soils. In a multiple linear regression, organic carbon content and the mean grain size of the eroded soils explained 35% of the variation in 137Cs content. This relationship suggests that both organic carbon and 137Cs are removed by erosion, with erosion being more prevalent on soils with a finer sand fraction. Clay and silt contents do not vary with depth in the near-surface horizons of the eroded sandy soils, hence it is likely that wind erosion strips the entire surface horizon with its 137Cscontent, rather than selectively winnowing fine material