Semiarid steppe ecosystems account for large terrestrial areas and are considered as large carbon (C) sinks. However, fundamental information on topsoil sensitivity to grazing is lacking across different spatial scales including the effects of topography. Our interdisciplinary approach considering soil chemical, physical, and vegetation properties included investigations on pit scale (square-metre scale), plot scale (hectare scale), and the scale of a landscape section (several hectares). Five different sites, representing a grazing intensity gradient, ranging from a long-term grazing exclosure to a heavily grazed site were used. On the pit scale, data about aggregate size distribution, quantity of different soil organic carbon (SOC) pools, SOC mineralisation, hydraulic conductivity and shear strength was available for topsoil samples from representative soil profiles. Spatial variability of topographical parameters, topsoil texture, bulk density, SOC, water repellency, and vegetation cover was analysed on the basis of regular, orthogonal grids in differently grazed treatments by using two different grid sizes on the plot scale and landscape section. On the pit scale, intensive grazing clearly decreased soil aggregation and the amount of fresh, litter-like particulate organic matter (POM). The weak aggregation in combination with animal trampling led to an enhanced mineralisation of SOC, higher topsoil bulk densities, lower infiltration rates, and subsequently to a higher risk of soil erosion. On the plot scale, the effects of soil structure disruption due to grazing are enhanced by the degradation of vegetation patches and resulted in a texture-controlled wettability of the soil surface. In contrast, topsoils of grazing exclosures were characterised by advantageous mechanical topsoil characteristics and SOC-controlled wettability due to higher POM contents. A combined geostatistical and General Linear Model approach identified topography as the fundamental factor creating the spatial distribution of texture fractions and related soil parameters on the scale of a landscape section. Grazing strongly interfered with the topography-controlled particle relocation processes in the landscape and showed strongest effects on the aboveground biomass production and biomass-related soil properties like SOC stocks. We conclude that interdisciplinary multi-scale analyses are essential (i) to differentiate between topography- and grazing-controlled spatial patterns of topsoil and vegetation properties, and (ii) to identify the main grazing-sensitive processes on small scales that are interacting with the spatial distribution and relocation processes on larger scales.