Wind erosion preferentially removes the finest carbon- and nutrient-rich soil fractions, and consequently its role may be significant within terrestrial carbon (C) cycles. However, the impacts of wind erosion on soil organic carbon (SOC) redistribution are not considered in most carbon cycle models, or within national carbon accounting schemes. Current estimates of the loss of SOC to wind erosion do not consider the preferential removal and enrichment of SOC in dust emissions, and consequently underestimate the significance of carbon dust emissions for the carbon cycle and carbon accounting. We present a process-based approximation of SOC enrichment within the Computational Environmental Management System (CEMSYS v5) Australian national wind erosion model. It enabled the prediction of carbon dust emission at a 50 km spatial resolution across Australia each month from 2000–2011. Our results show that the total dust emission for Australia is 118 Mt yr-1, with carbon dust emissions (<22 µm) in the order of 1.59 Mt yr-1. Rangelands are found to produce ~84% of Australia’s carbon dust emissions (1.34 Mt yr-1), while agricultural lands produce 0.11 Mt yr-1. However, agricultural lands produce larger carbon dust emissions per km2 than rangelands. The long-distance transport of SOC by wind is expected to result in large mineralisation rates due to (e.g.) photochemical oxidation, and the release of up to 5.9 Mt CO2-e yr-1. These results on the loss of SOC in dust, and the release of CO2 during long-distance transport, demonstrate the need for aeolian processes to be given greater significance in terrestrial carbon cycling, and for them to be incorporated into carbon accounting systems.