Atmospheric CO2 reconstructions are currently available from direct measurements of air enclosures in Antarctic ice and, alternatively, from stomatal frequency analysis performed on fossil leaves. A period where both methods consistently provide evidence for natural CO2 changes is during the 13th century ad. The results of the two independent methods differ significantly in the amplitude of the estimated CO2 changes (10 ppmv ice versus 34 ppmv stomatal frequency). Here, we compare the stomatal frequency and ice core results by using a firn diffusion model in order to assess the potential influence of smoothing during enclosure on the temporal resolution as well as the amplitude of the CO2 changes. The seemingly large discrepancies between the amplitudes estimated by the contrasting methods diminish when the raw stomatal data are smoothed in an analogous way to the natural smoothing which occurs in the firn.