The use of magnetic susceptibility to measure long-term soil redistribution

Several studies have documented the severity of recent soil erosion on the Canadian prairies where cultivation started about a century ago. Little quantitative information is available on erosion before 1960. This study attempts to quantify post-and pre-1960 soil erosion in a small cultivated basin near Saskatoon, Saskatchewan, Canada, by measuring (137)Cs and magnetic susceptibility (chi) distribution with depth. Soil cores were collected along six transects (three across closed depressions and three across the drainage channel) in the cultivated field, and one transect across an uncultivated depression. The cores were sliced into 3-cm layers and the soil analyzed for (137)Cs, chi, and organic and inorganic C. High variability in chi with depth in eroding areas (as indicated by (137)Cs) made it impossible to use chi to quantify past soil losses in these locations. However, these eroding upper and middle slope positions have a much higher chi than lower slope areas where soil deposition occurs and where the variability in chi with depth could be used to estimate soil deposition. Estimating soil deposition from the chi vs. depth profiles was more successful in the closed depressions than in the drainage channel, where the chi profiles may reflect the variable source areas of the materials rather than the pedological conditions. The data indicated that soil deposition since 1960 has been about 30 to 50% of that prior to 1960. This suggests that soil deposition rates, and by implication, soil erosion rates, have been relatively steady since cultivation of these soils started, although there are clear indications that the spatial pattern of deposition has varied. (C) 1998 Elsevier Science B.V.