Land use change-induced, accelerated soil erosion is a main driver of soil organic carbon (SOC) loss. However, soil erosion and SOC loss are not linked linearly because SOC-rich topsoil is eroded at the initial stages of degradation while horizons with lower SOC content are eroded later, but often at higher rates. To assess the contribution of soil erosion to the changing global carbon cycle, it is essential to understand how SOC changes when soil erosion accelerates. On cropland, dynamic replacement of organic carbon (OC) limits the net loss of soil OC stocks and causes a small, but steady flux of C from the atmosphere to soil and onwards landscape sinks. Rangelands, on the other hand, may loose their entire OC pool as a result of soil erosion because soils loose their productivity quickly when affected by erosion, disabling the replacement of SOC. Most rangelands have experienced erosion and degradation during the 20 ^th century, i.e. at the same time when greenhouse gas emissions from burning fossil fuels also increased. Distinguishing the contributions of rangeland land use change from other sources is essential for understanding the human impact on the global carbon cycle and climate during the 20 ^th century. The use of small farm reservoirs has been widespread in the South African Karoo during the last century. Many reservoirs have silted up and thus potentially document the soil properties during the period of sediment accumulation. This study tests the potential to reconstruct 20 ^th century SOC and total nitrogen (TN) change in rangeland soils. Five reservoir sediment profiles were sampled and texture, total organic carbon (TOC), TN and ¹³⁷Cs of the samples were analysed and compared along the reservoir profiles. The results show that with one exception the oldest sediments contain more TOC and TN than the older ones. This indicates a loss of SOC and TN early after the dams were constructed in the 1920s and 1930s. These distinct changes also illustrate that the small reservoir sediments have the potential to reconstruct the impact of land use, soil erosion and associated SOC change in rangelands.