The effect of cultivation on the spatial variability of selected properties of an umbric horizon

The main objective of this paper is to assess the spatial variability of some soil properties following two contrasting soil uses in Galicia, northwest Spain. Adjacent plots, one with natural vegetation and the other cultivated, were marked on regular square grids with 2-m spacing. Soil samples were collected from each sampling point at an acid, rich in organic matter, umbric horizon and analyzed for routine soil chemical properties and particle size distribution. Most of the soil properties analyzed had normal distributions on both land uses. Nitrogen, phosphorus, cation exchange capacity (CEC) and pH in H2O for the natural vegetation and silt content for the cultivated field were the exceptions. Organic carbon content and CEC of the cultivated field were higher. Gravel content also increased with cultivation, because of mixing through deep plowing. The cultivated soil was found to be much more homogeneous than the adjacent soil with natural vegetation. The naturally vegetated site presented a high and statistically significant correlation between organic matter content and CEC, which was not found for the cultivated site, indicating that most of the exchange capacity of this soil is on the organic matter and can be depleted very easily after clearing and cultivation. Organic matter, CEC, pH in H2O and phosphorus exhibit spatial dependence at the sampled scale and their experimental semivariograms were adjusted to a spherical model. Gravel content for the cultivated plot had a continuous semivariogram as compared to a pure nugget effect for the plot with natural vegetation. The pattern of spatial variability of some properties was found to be different for the two land uses, indicating that long-term cultivation changed the natural point to point continuity. The nugget effects for organic matter content and CEC were smaller on the cultivated plot, suggesting that agricultural soil use improved the continuity at close distances of these attributes. The cross-semivariogram between organic matter and CEC for the plot with natural vegetation was also adjusted to a spherical model with lower nugget effect than single semivariograms. The maps obtained with kriging illustrate much less diversity for the cultivated soil as well as the CEC dependence on organic matter content for both, cultivated and never-cultivated soils.