The δ13C values of pedogenic (soil-formed) calcite preserved in the sedimentary record have been used to estimate atmospheric pCO2 using the paleosol calcite paleobarometer. A fundamental assumption for applying this paleobarometer is that atmospheric CO2 concentrations have a direct influence on the measured pedogenic calcite δ13C values as a result of open-system exchange between atmospheric and soil-respired CO2. Here we address the timing of calcite precipitation in relation to the soil saturation state and soil-atmosphere connectivity in a modern Vertisol (smectitic, clay-rich soil, seasonally saturated) in Brazoria County, Texas, U.S.A. Luminescent phases of calcite growth, under cathodoluminescence microscopy, have more negative δ13C values (δ13C = −11.1‰ VPDB ± 0.78 1σ) than the non-luminescent phases (δ13C = −2.53‰ VPDB ± 1.41 1σ). The luminescent phase of calcite formed during the water-saturated portion of the year, thereby minimizing the incorporation of atmospheric CO2, and negating its use for pCO2 estimations. The non-luminescent phase formed during the well-drained portion of the year when atmospheric CO2 mixed with soil-respired CO2 and is therefore useful for pCO2 estimation. From these results we present a model to independently test the saturation state of a paleosol at the time of pedogenic carbonate precipitation. Finally, we calculate soil-respired CO2 concentrations that are an order of magnitude lower than those that are typically assumed in the soil-carbonate paleobarometer equation.