Many researchers are interested in the variability of root-respired δ13CO2 as an indication of linkages between belowground plant respiration and canopy processes. Most studies in this area have, however, relied upon the assumption that temporal variability of total soil respired δ13CO2 reflects autotrophic soil processes, but in fact few supporting measurements of purely autotrophic soil respiration (partitioned from total soil respiration) are available. Here we use a combination of physical and isotopic partitioning methodologies to track the variability in autotrophic and heterotrophic soil δ13CO2 at five sites in Eastern Canada during a very dry growing season. Three dimensional modeling of soil isotopic transport dynamics in the static sampling chambers allow us to constrain measurement bias and to eliminate non-steady-state effects as a potential driver of observed variability. We provide experimental results that support a pivotal assumption made in prior interpretations of soil δ13CO2 dynamics: we observed minimal isotopic variability in soil microbial δ13CO2 efflux, but appreciable temporal variability in root-respired δ13CO2 at sites where near drought conditions were observed, suggesting that isotopic discrimination is likely linked to seasonal variations in transpirational demand.