We estimated regional fluxes of carbon dioxide (CO2) using mixing ratios measured in a tallgrass prairie in central Kansas, USA over 3 yr (2002–2004). Glass flasks were used to collect whole air samples in the midafternoon for determining CO2 mixing ratios and their carbon isotopic composition. Regional CO2 fluxes were calculated assuming atmospheric boundary layer (ABL) approaches an equilibrium state on a monthly basis. CO2 mixing ratios derived from the marine boundary layer data were used as a proxy to represent those in the free troposphere, which allowed for determining a boundary layer CO2 gradient primarily resulting from surface exchange. We estimated temporal changes in the ABL height for this region on a monthly basis (600–1700 m asl for a 5-yr average between 1997 and 2001) from European Center for Medium-Range Weather Forecasts (ECMWF) model data. Accordingly, we estimated the rate of entrainment (flux density) by interpolating NCAR/NCEP reanalysis data to the estimated ABL height. Our study differentiates from previous studies in several aspects: (1) we used flask-based mixing ratio measurements; (2) only discrete midday CO2 mixing ratio data were used to construct weekly CO2 gradients between free troposphere and the ABL and (3) we propose a new means for estimating monthly values of vertical transport. Modelled regional CO2 fluxes were compared to net ecosystem exchange (NEE) of CO2 fluxes measured by eddy covariance method. Assuming negligible vertical CO2 gradients between mid-ABL and the surface layer and with no correction applied, calculated NEE showed a general agreement with measured NEE fluxes throughout the growing season. Using CO mixing ratio data, we show that fossil fuel burning contributed negligible CO2 fluxes in summer but partially explained the discrepancy between modelled regional CO2 fluxes and measured NEE in winter. This wintertime fossil fuel input was consistent with carbon isotope measurements of CO2. We demonstrate in this study that CO2 mixing ratios subsampled at midday in the surface layer can be used to gain insights into regional CO2 flux exchange in the U.S. Great Plains area.