Spatial distribution and content of soil organic matter in an agricultural field in eastern Canada, as estimated from geostatistical tools

Soil erosion induces soil redistribution within the landscape and thus contributes to the spatial variability of soil quality. This study complements a previous experimentation initiated by the authors focusing on soil redistribution as a result of soil erosion, as indicated by caesium-137 (Cs-137) measurements, in a small agricultural field in Canada. The spatial variability of soil organic matter (SOM) was characterized using geostatistics, which consider the randomized and structured nature of spatial variables and the spatial distribution of the samples. The spatial correlation of SOM (in percentages) patterns in the topsoil was established taking into account the spatial structure present in the data. A significant autocorrelation and reliable variograms were found with a R-2 >= 0.9, thus demonstrating a strong spatial dependence. Ordinary Kriging (OK) interpolation provided the best cross validation (r(2) = 0.35). OK and inverse distance weighting power two (IDW2) interpolation approaches produced similar estimates of the total SOM content of the topsoil (0-20 cm) of the experimental field, i.e. 211 and 213 tonnes, respectively. However, the two approaches produced differences in the spatial distribution patterns and the relative magnitude of some SOM content classes. The spatialization of SOM and soil redistribution variability as evidenced by Cs-137 measurements is a first step towards the assessment of the impact of soil erosion on SOM losses to recommend conservation measures. Copyright (C) 2009 John Wiley & Sons, Ltd.