This research focuses on a semi-arid, agricultural area in Central Spain near Madrid, in which airborne hyperspectral images have been obtained. Small-scale soil erosion features are exposed at the surface as a consequence of human induced soil erosion derived mainly from tillage practice. Such features are associated with different soil horizons and rock outcrops with contrasted physical and chemical characteristics. Results show that the identification and mapping of different soil surface horizons linked to soil erosion and depositional stages can be achieved over selected test sites based on the spectroscopy data at high spatial resolution. Linked with field validation data and geomorphological analyses, the spatial mapping of the soil erosion and depositional stages is consistent with the soil erosion models implemented for the region. Preliminary multiscale analyses at 3m, 6m, and 30m show the effect of increasing spatial mixing in the field-of-view of the sensor due to the variability at small scale of the different soil horizons representing the surface topsoil.