Increasing soil carbon content via agricultural practices not only enhances the production potential of the land, but also counteracts rising atmospheric CO₂ levels. When predicting production systems’ effects on soil carbon, quantifying CO₂ efflux derived from live roots is of particular importance as it is a through-flux and does not signify depletion of soil carbon. This field study aimed to measure and compare soil CO₂ emissions derived from roots in annual and perennial agroecosystems. We used periodic 48-hour shading over two growing seasons to estimate root growth-derived CO₂ in continuously grown maize (CC) with grain and 50% stover harvested each year, unfertilized reconstructed tallgrass prairie (P), and the same prairie grown with spring nitrogen fertilization (PF), both which had biomass harvested post-frost. In CC, P, and PF root-derived CO₂ contributed to 28, 31, and 30% of each crop’s respective growing season cumulative CO₂ emissions in 2012, and 19, 24, and 28% in 2013, respectively. Season-cumulative root-derived CO₂ was not proportional to end-of-season belowground biomass (BGB): P had nearly twice the BGB of PF, but their cumulative root-derived fluxes were not significantly different in either year. A significant proportion of soil CO₂ emissions is derived from roots, making it a critical process to consider when comparing or modeling soil emissions of cropped or prairie soils. Using BGB alone may not be a useful proxy for estimating root contributions.