Conventionally, measurements of carbon isotopes in atmospheric CO<sub>2</sub> (δ<sup>13</sup>CO<sub>2</sub>) have been used to partition fluxes between terrestrial and ocean carbon pools. However, novel analytical approaches combined with an increase in the spatial extent and frequency of δ<sup>13</sup>CO<sub>2</sub> measurements allow us to conduct a global analysis of δ<sup>13</sup>CO<sub>2</sub> variability to infer the isotopic composition of source CO<sub>2</sub> to the atmosphere (δ<sub>s</sub>). This global analysis yields coherent seasonal patterns of isotopic enrichment. Our results indicate that seasonal values of δ<sub>s</sub> are more highly correlated with vapor pressure deficit (<i>r</i> = 0.404) than relative humidity (<i>r</i> = 0.149). We then evaluate two widely used stomatal conductance models and determine that the Leuning Model, which is primarily driven by vapor pressure deficit is more effective globally at predicting δ<sub>s</sub> (RMSE = 1.6‰) than the Ball-Woodrow-Berry model, which is driven by relative humidity (RMSE = 2.7‰). Thus stomatal conductance on a global scale may be more sensitive to changes in vapor pressure deficit than relative humidity. This approach highlights a new application of using δ<sup>13</sup>CO<sub>2</sub> measurements to validate global models.