Soil C inventory assessments indicate data gaps regarding the spatial variability of C stocks as affected by landforms in Southern Brazil. This study aimed at assessing soil C variability along toposequences on very clayey Oxisols farmed for 50 years. The study area in the Atafona watershed is representative of a major grain production area of Brazil, where the prevailing cropping system comprises soybeans (summer) and small grains(winter) under continuous no till.

Five toposequences from 400 to 1,200 m were selected in adjacent fields. Soil samples (0-20cm) were obtained from summits to toeslopes for each toposequence. Sampling avoided depositional areas at field edges where other soil classes occur. Greater C stocks (47.8 Mg ha^-1) were observed in toeslopes in comparison to summits and backslopes, in which C stocks were remarkably similar (44.4 and 44.1 Mg C ha^-1, respectively). These larger stocks at toeslope positions could be primarily ascribed to greater moisture, and associated higher biomass production. However, legacy contribution of C-enriched sediments transported from upper to lower landscape positions in the past must be considered. High erosion due to multiple tillage operations and no conservation practices were common throughout the 1970s and 1980s in this region.

Our results are in agreement with previous coarser-scale soil C stock inventories. These clayey Oxisols present remarkably small variations in C stocks of upper landscape positions (summits and backslopes) when double-cropped with soybeans and small grains under no tillage for long periods. Given the generally much larger area covered by these lardforms in relation to depositional areas, the latter contribute marginally to total C stock calculations. Landscape segmentation with finer scale soil maps and farm field elevation profiles that distinctly delineate toeslopes (and associated soil class transitions) could be instrumental in devising optimal soil sampling protocols to assess C stocks and C stock changes in these soils with greater accuracy.